Political Activism -in Pollock G. and Ellison M. (2014) 'Measuring Participation', European...

MYPLACE August 2014

MYPLACE: FP7-266831 www.fp7-myplace.eu Deliverable 4.6 Europe wide thematic report of 431

____________________________________

____________________________________

MYPLACE (Memory, Youth, Political Legacy And Civic Engagement)

Grant agreement no.: FP7-266831

WP4: Measuring Participation

Deliverable 4.6: Europe-wide thematic report

Editors Mark Ellison, Gary Pollock Version Final v5.1 Date 12th August 2014 Work Package WP4: Measuring Participation Deliverable 4.6: Europe wide thematic report Dissemination level PU: Public WP Leaders MMU, CRRC Deliverable Date June 2014 Document history Version Date Comments Modified by V5.0 7th August 2014 Final draft version ME V5.1 12th August Final version for submission GP

http://www.fp7-myplace.eu/index.php

MYPLACE August 2014


Contents:

Page:

Authors and Acknowledgements 3

Chapter 1: Key findings 6

Chapter 2: Introduction 12

Chapter 3: Survey Design and Methodology 14

Chapter 4: Political Interest 48

Chapter 5: Attitudes and Trust 97

Chapter 6: Political Activism 127

Chapter 7: Civic Engagement 164

Chapter 8: Effective Political Action 178

Chapter 9: Citizenship 207

Chapter 10: Social Networks, Gender and Sexuality 251

Chapter 11: Religion 277

Chapter 12: Attitudes Towards Minority Groups 293

Chapter 13: Democracy 331

Chapter 14: History and Memory 365

Chapter 15: European Issues 388

Chapter 16: Summary 404

Appendix A: Guidance Document 428

MYPLACE August 2014


Authors and Acknowledgments

Chapter 1: Key Findings

Authors: Mark Ellison and Gary Pollock

Chapter 2: Introduction

Authors: Mark Ellison and Gary Pollock

Chapter 3: Survey Design and Methodology

Authors: Tina Zurabishvili, Tamuna Khoshtaria, Koba Turmanidze and Bret Barrowman

(Caucasus Research Resource Centers, Georgia)

Chapter 4: Political Interest

Authors: Marti Taru and Ladislav Machacek

We would like to thank Inta Mierina for her help with multilevel regression and Mark Ellison

for feedback and help with formatting.

Chapter 5: Attitudes and Trust

Authors: Renata Franc and Vanja Međugorac

We would like to thank the young people at Peščenica and Podsljeme in Zagreb who

participated in survey.

Chapter 6: Political Activism

Authors: Roger Soler i Martí and Mariona Ferrer-Fons (Universitat Pompeu Fabra,

Barcelona) and Augusta Correia, Nuno de Almeida Alves and David Cairns (CIES-IUL/ISCTE-

IUL, Lisboa)

We would like to thank;

Portugal: fieldwork contractor: TNS

UPF: fieldwork contractor: GESOP and the councils of Vic and Sant Cugat

Chapter 7: Civic Engagement

Authors: Vesa Puuronen, Kari Saari, Samuli Neuvonen and Niko Eskelinen (Universty of

Oulu, and University of Eastern Finland, Finland) and Klaus Levinsen, (University of Southern

Denmark)

Chapter 8 Effective Political Action

Authors: Tina Zurabishvili, Tamuna Khoshtaria, Koba Turmanidze and Bret Barrowman

(Caucasus Research Resource Centers, Georgia)

MYPLACE August 2014


Chapter 9: Citizenship

Authors: Olga Elkina, Head of Marketing and Social Research, Natalia Goncharova, Deputy

Director, Scientific Research Center "Region", Senior Lecturer, Ulyanovsk State University

and Yana Krupets, Deputy Director, Centre for Youth Studies, Higher School of Economics, St

Petersburg.

The authors would like to thank: Olga Echevskaya (Novosibirsk State University, Russia),

Inta Mierina (Daugavpils University, Latvia), Tinatin Zurabishvili (Caucasus Research

Resource Centers program, Georgia), Gary Pollock (Manchester Metropolitan University,

United Kingdom), Hilary Pilkington (Manchester University, United Kingdom), Daria Krivonos

(Centre for Youth Studies, Saint-Petersburg University Higher School of Economics, Russia),

Dmitry Omelchenko (Centre for Youth Studies, Saint-Petersburg University Higher School of

Economics, Russia), Rowenna Baldwin (Manchester Metropolitan University, United

Kingdom)

We would also like to thank our respondents and our interviewers.

Chapter 10: Social Networks, Gender and Sexuality

Authors: Britta Busse, Alexandra Hashem-Wangler, Jochen Tholen (University of Bremen)

We would like to thank the Public Statistical Office of the state of Bremen for the careful

and professional execution of the WP4 fieldwork and the excellent cooperation.

Chapter 11: Religion

Authors : Irini Chiotaki-Poulou, Alexandra Koronaiou and Alexandros Sakellariou.

We would like to thank Emmanouella Kostopoulou, Project Leader of the subcontracting

company for her valuable help during the survey.

Chapter 12: Attitudes towards Minority Groups

Author: Inta Mierina is a lead researcher in the Daugavpils University, Latvia

We are grateful to the Institute of Philosophy and Sociology, University of Latvia, particularly

Ilze Koroļeva and Aleksandrs Aleksandrovs, for their help with preparation of the survey and

data collection in Latvia.

Chapter 13: Understandings of Democracy

Authors: Mark Ellison, Haridhan Goswami and Gary Pollock (Manchester Metropolitan

University)

We would like to thank Fieldwork Supervisor Marius Guderjan, the Herbert Museum,

Coventry and the young people of Coventry and Nuneaton who participated in the survey.

MYPLACE August 2014


Chapter 14: History and Memory

Authors: Istvan Muranyi, Zoltan Berenyi

We would like to thank István Vas (Adviser, Mayor's Office Ózd) and Balázs Kanczler

(Sociologist, Sopron) for their assistance in the survey work.

Chapter 15: European Issues

Authors: Mark Ellison and Gary Pollock, with extracts from various chapters.

General Acknowledgement

We would like to thank the young people across Europe who participated in this study, all

the fieldworkers and the various public, private and voluntary organisations who supported

the fieldwork. We hope to have done justice to the collective efforts of so many people in

our analyses. Without funding under the EU’s Framework 7 programme none of this work

would have been possible.

MYPLACE August 2014


1.0 Key Findings

These key findings are based on the the responses of almost 17,000 young people, aged 16-

25 in 30 locations across 14 countries in Europe who took part in the MYPLACE survey

(Workpackage 4: Measuring Participation).

The case study design, using two contrasting locations per country (four in Germany), allows

comparison both within and between countries. The findings detailed in this report

demonstrate that employing this methodology reveals that there are geographical patterns

which vary according to the subject of the analysis. It is often the case that country specific

similarities between locations are accompanied by significant between country

differentiation. However, this should not be regarded as evidence of a ‘national average’ as

our research locations are not selected to represent each nation and while any association

between the country specific findings reported here and nationally representative results

may be noted, this is not an intended feature of our research design. Indeed, there are

many instances of significant within country differences between the research locations

making it possible to identify greater similarities between locations in different countries.

An illustrative example of this concerns respondents’ views of the importance of language

competence to being a citizen of a given country. On this question, respondents in Tartu

(Estonia) are the most likey to say that ‘to speak the national language’ is important (96%)

while those in the Narva area, the other Estonian location but located in the east of the

country close to the border with Russia and with a high proportion of ethnic Russians, are

much less likely to say that it is important (54%). Similar differences also occur in Latvia,

with 84% of respondents in the Agenskalns district of Riga saying it is important as opposed

to 48% in the Forstate & Jaunbuve, districts of Daugavpils, a city close to the Russian border

also with a high proportion of ethnic Russians. The next highest within country variance on

this question is in Hungary where 92% of respondents in Ozd, a town close to the border

with Slovakia, state that national language competence is important for citizenship

contrasting with 81% of respondents in Sopron which is close to the Austrian border.

Interest in Politics

Across all locations, more young people reported that they are ‘not interested’ (58%) rather

than are ‘interested’ in politics (42%). The levels of interest were differentiated at the

country and wider regional level; young people were more likely to be interested in politics

in eastern Germany, Greece and Spain, but less likely in Croatia, Latvia, Finland and Estonia.

Employment, housing and the environment are the three issues that are of greatest interest

for these young people. The European Union, immigration and LGBT rights, in contrast,

scored lowest in terms of the level of political interest.

MYPLACE August 2014


Television and the Internet are the two modes of information used most by young people to

obtain information about politics and public affairs.

Attitudes and Trust

Young people feel very much detached from politicians and tend to have a negative view of

them. Only 22% of our respondents agree with the statement: ‘politicians are interested in

young people like me’ whereas 60% agree with the statement ‘politicians are corrupt’ and

69% agree with the statement ‘the rich have too much influence in politics’. When

examined regionally, a majority of the Danish respondents actually agree that politicians are

interested in them whereas over 80% of our Greek respondents disagree.

On a 0 to 10 scale, young people tend not to trust the Prime Minister with an average value

of 4.3, parliament (4.5) and political parties (3.8) in their countries although the findings also

reveal country and regionally based patterns. The least trusting responses were

characteristic for young people from Greece, Croatia, Spain and Hungary, whereas Finland,

Georgia, Denmark, eastern and western Germany have levels of trust at the midpoint of the

scale.

Political Activism

Of those young people eligible to vote in national elections, 71% actually did at the last

opportunity and 17% made a conscious decision not to vote. The remaining 11% reported

that they would have voted but were unable to. Locations with voting levels in excess of

80% are; Podsljeme, both Danish locations, Kuopio (Finland), both eastern German

locations, Agenskalns (Latvia) and Vic (Spain). By far the lowest level of voting was in

Nuneaton (UK) where only 39% of eligible voters chose to do so, with 43% of eligible voters

deliberately choosing to stay away from the polls.

More than half of respondents (59%) state that they do not feel close to any political party.

There are large variations in the level of political activism across our locations. The global

level of non-electoral participation (an index based on 20 different political activities)

showed large differences between the 30 locations in the study. As an example, both

locations in Hungary represent a level of participation 25 times lower than some locations in

eastern Germany or Spain.

In locations that have experienced a greater impact of the current financial crisis, young

people participate less in global terms but tend to participate more through protest actions.

MYPLACE August 2014


Civic Engagement

Young people have extremely low levels of civic engagement. A composite index of eleven

civic and political organisations1 shows the overall average to be 0.3 on an eleven-point

scale ranging from 0 (no engagement) to 11 (high engagement). There is considerable

variation in civic engagement across locations and countries . Young people from the Nordic

countries (Denmark and Finland), have the highest levels of membership of civic and

political organisations, while young people from Mediterranean (Portugal, Spain and

Greece) and post-socialist countries (Hungary, Slovakia, Georgia and Latvia) have the lowest

levels.

Effective Political Action

On a 0 to 10 scale, voting in elections received the highest mean score for effectiveness of

political action (6.9), followed by gaining publicity through media exposure (5.9) and being

active in a political party (5.4). Participating in violent and illegal protest activities are

universally regarded as the least effective with scores of 2.9 and 3.5 respectively.

Young people in locations in Georgia tend to rate the range of non-violent and violent

political actions as relatively effective, while young people in Russia, Hungary, Croatia and

Slovakia tend to be relatively pessimistic about the effectiveness of political action.

Citizenship

Those who are most proud of their citizenship status2 are the young people from locations

in Georgia, Finland, Denmark and the UK with levels of pride around 90%. Approximately

60% of the young people from Vic and Sant Cugat (both locations are in Catalonia) state that

they are not proud of their citizenship. Other locations where levels of pride in citizenship

were less than 70% are: the Narva area (Estonia), Bremen (western Germany) Jena and

Rostock (both in eastern Germany), and Forstate & Jaunbuve (Latvia).

1 The organisations included in the index are: ‘political parties or youth sections of political parties’; ‘religious or church organisations or religious youth organisations’; ‘trade unions or youth organisations of trade unions’; ‘national or local youth parliaments’; ‘environmental organisations’; ‘animal welfare groups’; ‘peace organisations’; ‘human rights organisations’; ‘national or cultural organisations (for instance diaspora or ethnic minority organisations)’; ‘women’s organisations’ ; and ‘anti-globalisation organisations or movements’ 2 Our analysis is sensitive to nuances across our research locations where nationality, ethnicity and citizenship

have specific local connotations. The large ethnically Russian population in the Narva area in Estonia as well as

the Forstate and Jaunbuve districts in Latvia are particularly important in this respect. In addition, Coventry in

the UK is particularly ethnically diverse. Our analysis therefore uses a measure of ‘identity’ based on citizen,

nationality and ethnic status.

MYPLACE August 2014


Respondents from locations in Greece, Croatia , Slovakia and Portugal tend to support

limiting the import of foreign products in order to protect the national economy. In

contrast, locations in Denmark as well as eastern and western Germany tend not to support

this policy. The idea that the ‘country should follow its own interests even if this leads to

conflicts with other nations’, is supported by the majority of respondents in locations in

Georgia, Greece, Hungary and Portugal.

Social networks, gender relationships and sexuality

Levels of trust in people were highest in locations in Finland, western Germany and

Portugal, while Greek, Russian, Slovakian and Georgian locations had the lowest mean

values for trust in people.

In terms of attitudes towards women and homosexuality in society, Denmark, Finland, Spain

and Germany tend to have the most egalitarian responses. The least egalitarian views were

expressed in locations in Georgia, Greece, Slovakia, Latvia and Russia.

Women`s access to abortion is supported most strongly in locations in eastern Germany,

Spain and Estonia. Young people in locations in Russia, Greece and Georgia are the most

likely to believe that women should not have access to abortion.

Religion

Levels of religiosity are high in a cluster of locations including Georgia, Greece, Croatia and

Russia. Locations in eastern Germany, Spain and Denmark, on the other hand, are among

those showing the lowest levels of religiosity.

Locations in Georgia and Greece have higher levels of agreement to the statement that

‘there is only one true religion’, as opposed to eastern Germany that uniformly has the

lowest levels of agreement. Locations in Spain, Greece and western Germany tend to agree

that religious leaders should have no influence on government decisions.

Attitudes towards minority groups

Young people in Western European locations, especially eastern Germany, are least likely to

have negative views towards different minority groups and to advocate limiting the access

of migrants to a range of resources. Young people in post-socialist locations, along with

Greece, tend to express more negative views and are more likely to be exclusionist towards

migrants.

MYPLACE August 2014


Young people in locations in Greece are much more likely than others to have xenophobic

attitudes, whereas eastern and western German locations are clustered at the bottom of

the xenophobia scale.

Locations in Hungary, Russia and Georgia are at the top of the welfare chauvinism and

exclusion scale; in these locations young people favour limiting access by migrants to a

range of resources.

Democracy

Young people’s satisfaction with democracy in their country tends to be around the middle

of a 0-10 scale across all locations although a number of clear clusters emerge. Locations in

Denmark, Finland, eastern and western Germany have higher average levels of satisfaction

with democracy whereas mediterranean countries including Greece, Spain and Portugal

have lower average levels.

When asked about the desirability of different types of political systems, respondents from

locations in Denmark and eastern Germany tend to have more positive views towards

democratic political systems, whereas, locations in Russia, Latvia and Slovakia tend to be

less positive.

Young people in locations in Georgia, Croatia, the UK and Portugal tend to have more

positive attitudes towards non-democratic political systems3. Locations in Germany

(western and eastern), Finland and Spain are more likely to be negative.

History and Memory

Young people are interested in recent history, with 29% stating that they are very interested

and 56% a little interested. Interest in recent history was highest in locations in Greece,

Georgia, and Germany (eastern and western).

The reported importance of different historical events shows that respondents consider

World War II (85%) and membership in the European Union (84%) to be ranked highest

from their countries’ perspectives among the listed events and periods. The NY terrorist

attacks of September 9th 2011 (41%) and the Holocaust (52%) are less likely to be seen as

important. There is, however, massive location variance in these figures, for example the

importance of the holocaust varies from a low of 18% in Lieksa/Nurmes (Finland) to 91% in

Jena (eastern Germany). Locations where the importance of the holocaust is greater than

60% are: Nuneaton and Coventry (both in the UK), Vyborg (Russia), Sopron (Hungary),

Rostock, Jena, Bremerhaven and Bremen, i.e. all four German locations.

3 Attitude to ‘Having a strong leader who is not constrained by parliament’ and ‘Having the army rule’.

MYPLACE August 2014


European Issues

Interest in the European Union is high in Mediterranean locations (Portugal, Spain and

Greece) and locations in Germany (eastern and western). The highest scores, however, are

from both Georgian locations, the only ones to score averages higher than 7 on a 0-10 scale

where the overall average is 5.5. The lowest scores are from Lieksa and Nurmes (Finland)

with 4.1, Nuneaton (UK), with 4.2, and both Russian locations which each have a score of

4.0.

When asked about whether or not membership of the EU benefits their country, 49%

agreed with this statement4. The majority of respondents agree that EU membership

benefits their country in all four German locations and in both Spanish ones. However,

young people in Greece, Hungary and Rimavska Sobota (Slovakia) are far more lkely to

disagree.

On a 0 to 10 scale the average level of trust in the European Commission is 5.2, higher than

average trust in national parliaments at 4.5. The highest average levels of trust in the EC are

in both Georgian locations, Telavi (6.6) and Kutaisi (6.1) as well as the Finnish location of

Kuopio (6.3). The lowest average levels of trust were reported in both Greek locations, New

Philadelphia (3.7) and Argyroupouli (3.5) as well as the Russian location of Kupchino (4.3).

4This question was not asked in Russia or Georgia. A modified version was asked in Croatia where EU

membership was expected to be granted in July 2013.

MYPLACE August 2014


2.0 Introduction

This report is based upon the data analysis from a representative sample of young people

(16,935 questionnaires achieved) in the 30 MYPLACE research locations in 14 countries

across Europe.

2.1 Aims

The aim of the MYPLACE survey and this report in particular, is to present a general Europe

wide thematic analysis of young people’s views, attitudes and behaviour towards a range of

social and political issues. These attitudes and behaviours are analysed across the 30

research locations in 14 countries across Europe. The findings are arranged under the

following themes: political interest, political participation, citizenship, social networks,

gender and sexuality, religion, minority groups, understanding of democracy, history and

memory and European issues.

2.2 Purpose and scope of this report

This report is the final MYPLACE Work Package 4 deliverable and forms one of a series of

reports and papers from the survey component of the MYPLACE project. It is a key

deliverable for the European Commission and is a source document for future publications.

This report is aimed at a range of audiences including the European Commission, national

youth groups including the MYPLACE Youth Policy Advisory Groups (YPAGs), the academic

community and the young people who have participated in the survey.

The scope of this report is limited to the analysis of the 30 research locations and key socio-

demographics of the respondents. Each themed chapter is structured in the same way, and

contains the following sub-sections:

A review of substantive literature on the theme.

A brief summary of the fifteen individual country specific analyses, which were

produced as ‘Deliverable 4.5: A series of country specific analyses which highlight

local historical and cultural factors and which contrast the two locations sampled’.

MYPLACE August 2014


Comparative analysis of each of the 30 research locations on each question. When

there were good theoretical grounds to combine individual questionnaire questions

or items, factor analysis/Cronbach Alpha and further tests of robustness were used

to ensure these variables were suitable. Otherwise, individual questions or items

were analysed separately. Data are presented as a series of graphs with

supplementary analysis.

A multivariate analysis using one or more Multi-level models (MLM) (or in the case of

chapter 14 a multiple regression model). A multi-level model is favoured due to the

nested structure of the dataset, with locations within countries.

A summary of key findings from the chapter.

MYPLACE August 2014


3.0 Survey Design and Methodology

3.1 Questionnaire content

The final MYPLACE source questionnaire (in English) is presented in deliverable D4.1. The

questions cover the following topics: Political interest, political participation and civic

engagement (Section I); Social networks, gender relationships and sexuality (Section II);

Religion (Section III); Attitudes towards minority groups (Section IV); Understanding of

democracy, violence and human rights (Section V); History and memory (Section VI); and

Respondent’s socio-demographic profile (Section VII).

In order to ensure comparability between data collected by the MYPLACE partners, the

process of translation of the questionnaire has mainly used the ASQ (Ask the Same

Question) model. This allowed the data to be pooled for a trans-European analysis.

However, ASQ is not a direct translation. Unlike direct translations, ASQ does not ignore

cultural and linguistic nuances. Translation Tracking Forms (TTF) were used by all MYPLACE

partners, except the UK, to detail all issues in the wording of questions, instructions and

answer options, and to propose ways of solving these issues.

In respect to some questions, national variation was applied. For instance, level of political

knowledge was measured in questions 28_1 through 28_3 using a ‘functional equivalence’

approach, given that the aim of these questions was to measure political knowledge within

a national context, in order to produce a scale which differentiated levels of respondents’

knowledge.

In addition to Q28, national variation was required in the following questions:

Q7_3, where respondents’ trust towards the head of government was measured.

There are minor differences in how the question was worded by the national teams

(most partners asked about the Prime Minister, while several partners asked about

the President), but in all cases the question was asked about the top political figure

in the country. Hence, the answers have been aggregated in the merged data set

into a single variable irrespective of exactly how the question was worded.

In Q8 and Q11 participation in national and local elections was measured. The last

national and local elections were named by each partner.

In Q9, Q12 and Q15 respondents were asked to name the political party they voted

for (Q9, Q12) or felt closest to (Q15). Each partner provided a list of political parties

that participated in the corresponding elections (national or local).

MYPLACE August 2014


Q18_16 was an optional question about membership/participation in activities of an

‘other’ organisation (in addition to 15 organisations specified in earlier questions).

Only eastern German and western German teams used the opportunity to ask about

additional (16th and 17th) organisations, “youth fire brigade” (Q18_16) and “youth

club” (Q18_17).

Q23 was asked only to the respondents who were not citizens of the country where

the interview was conducted. Eleven partners used this variable: Croatia, Denmark,

Estonia, Greece, Hungary, Latvia, Portugal, Russia, Slovakia, Georgia, and the UK.

Q26_5 (assessment of the role of EU membership) was not asked in countries which

were not members of the European Union (Georgia and Russia).

In Q40_7 and Q40_8 the partners generated country-specific statements to measure

attitudes towards national and/or religious minority groups in their countries. In

accordance with the decision made during the MYPLACE project meeting in

Smolenice in May 2013, these country-specific variables were dropped from the

merged data set, but will be used for national-level analysis as appropriate.

Q51_9 (assessment of the importance of EU membership for the history of the

country) was only asked in EU member countries.

In questions Q51_10 through Q51_12 the partners had the opportunity to ask about

the importance of country-specific event(s) for the history of their country. In

accordance with the decision made during the MYPLACE project meeting in

Smolenice, these country-specific variables were dropped from the merged data set,

but will be used for national-level analysis as appropriate.

In Q58 (respondent’s ethnic group) ethnic groups relevant for each of the partner

countries were named.

In Q59 (respondent’s religion) religious groups relevant for each of the partner

countries were named.

Q60, Q73 and Q77 measured the highest level of education received by the

respondent, the respondent’s father, and the respondent’s mother respectively. The

answer options provided in the local language questionnaires represented

educational qualifications relevant for the partner countries.

The Croatian questionnaire had the following additional country-specific questions:

- Four additional statements measuring national identification:

o P26_6CRO “I believe Croatian policies are almost always morally correct”; o P26_7CRO “If I criticize Croatia, I do so out of love for my country”; o P26_8CRO “I support Croatia’s leaders even if I disagree with their actions”; o P26_9CRO “People should work hard to move this country in a positive direction”. Answer options were on a 5-point Likert scale.

- Q53_CRO “On which side was your family mainly during WWII?”

MYPLACE August 2014


with the following answer options:

o On the side of the Antifascists and Partisans, o On the side of the Croatian Home Guard, o On the side of Ustasha, o On some other side WRITE _____________, o On several different sides, o On none of the sides, o Do not know, o Refuse to answer.

- Q83 “During the Homeland war, …

P83_1. … were you or your family displaced (were forced to leave your home and

move to another part of the country)?

P83_2. … did you or your family have greater material losses?

P83_3. … did you or your family lose a family member?

P83_4. … did you or your family lose a close friend or neighbour(s)?”

Answer options included No, Yes, Do not know, Refuse to answer.

All these additional national variance questions were approved by the coordination team at

Manchester Metropolitan University and CRRC-Georgia with regard to their content and

their position in the questionnaire.

Additional questions and modifications to the eastern German questionnaire included:

In Q25, ‘Being a citizen of this country’ was changed to ‘being German.’

Q83 and Q83a, Place and Bundesland of birth were added to the questionnaire.

Three additional country specific questions were added in Greece, at the end of the

questionnaire, asking about respondents' feelings regarding their future life in Greece;

whether they have considered the possibility of moving abroad; and if so – for what reason.

3.2 Research locations

Each MYPLACE partner selected two contrasting locations to conduct the WP4 survey. The

D3.1 report provides detailed information about the rationale for the selection of survey

locations, which is summarized below.

Croatia

In Croatia, the WP4 survey was conducted in two districts of Zagreb: Podsljeme and

Peščenica. Zagreb is divided into 17 ‘city districts’. Podsljeme and Peščenica were selected

MYPLACE August 2014


as contrasting survey locations that are not adjacent and are very different in socio‐

economic as well as demographic terms. While Podsljeme is a rather quiet, mostly

residential and, in economic terms, prosperous area, Peščenica can be described as a

‘troubled’ part of the city. It is a poorer city district, mostly (but not exclusively) industrial,

and – at least to some extent – more ethnically mixed.

Several key factors were taken into account while selecting these locations. Firstly, based on

the available literature and empirical data on young people’s political and social

participation in Croatia, as well as the potential and propensity for radical ideas, it was

concluded that the most important factors influencing young people’s receptivity to radical

political agendas in Croatia were: (1) community segregation/cohesion (especially the

degree of contact with immigrants, ethnic and religious minorities); and (2) socio-economic

inequalities (including levels of education and unemployment, but also socio-economic

issues related to the 1991‐95 Homeland War, war veterans and war refugees). Secondly,

given the fact that different parts of the country had very different historical experiences, it

was decided that selecting two locations geographically remote from one another could

present a problem, as it would be difficult to make a meaningful comparison due to the

different issues of their historical past. While some regions were completely devastated

during The Homeland War of 1991‐5, others were left almost intact. Selecting two locations

within the same city would help to avoid this problem. Thirdly, considering that the WP4

survey required a sample representative for the youth living in each of the selected

locations (and, therefore, a robust sample frame), it was concluded that this information

would be most easily available for administrative units (e.g. ‘city districts’ or

‘municipalities’).

Denmark

In Denmark, the WP4 survey was conducted in the city of Odense. Odense has a population

of 170,327 which makes it the third largest city in Denmark and the main city of the island of

Funen. The fieldwork was conducted in Odense Centre and Odense East, which are two

relatively different parts of the city. Population of Odense Centre is socially mixed, though

with a preponderance of residents belonging to the middle class. Odense East is an area

which includes the suburb Vollsmose, a ghetto-like area with a high proportion of residents

with immigrant background. This area was a modern housing development project from the

1970s, and has about 10,000 inhabitants on its two square kilometres. Since the 1990s, this

area has become known throughout Denmark, due to a relatively high crime-rate and the

large density of people of foreign descent.

Estonia

MYPLACE August 2014


Selection of research sites in Estonia was guided by (1) the potential for understanding

youth socio-political (extremist) activism, and (2) ability to demonstrate a considerable

contrast between the locations in terms of the socio-economic characteristics and the

ethnic-cultural composition of the population.

In Estonia, the WP4 survey was conducted in two regions: Ida-Virumaa county in north-

eastern Estonia and Tartu county, in central-eastern Estonia. In Ida-Virumaa county,

interviews were carried out in three towns: Narva (66% of the sample of this location),

Sillamäe (8%) and Kohtla-Järve (26%). These settlements are populated largely by Russians

living in the socio-economic periphery, partly supporting anti-Estonian and pro-Soviet ideas.

A high percentage of this population is elderly.

In the central-eastern region, the survey was carried out in Tartu (81%) and in neighbouring

parishes. Population of Tartu is Estonian. They live in a large settlement, in a central

location, and are characterized by quite nationalistic ideas. This population has a high

percentage of young people.

Finland

The WP4 survey was conducted in eastern Finland, in the city of Kuopio (the city centre and

nearby areas of Niirala, Haapaniemi and Rönö) and the region of Pielinen Karelia (towns of

Lieksa and Nurmes). These fieldwork sites were selected on the basis of four criteria we

were looking for: (1) urban versus rural environment, (2) socio-economic differences

between the local population, (3) level of ethnic conflicts in the recent past, and (4) level of

support for the populist political party, The Finns.

Kuopio is the ninth largest city and an urban provincial centre in Finland. Every fifth person

living in the selected four areas is young (aged 15-24). In Kuopio, especially in the chosen

areas of the city, educational and income levels are higher than on average in the country.

The city has a relatively stable economy as it is a regional centre with quite a large and

attractive university, other institutions of higher education, industries, commercial, financial

and cultural services. However, the unemployment level in the whole Kuopio area is slightly

higher than the national average. Ethnically, the city is relatively homogeneous and the

number of ethnic conflicts (reported by the police) is low. In the selected areas of Kuopio,

the support for The Finns in the 2011 Parliamentary Election was relatively low.

Nurmes and Lieksa are relatively small towns in eastern Finland. These towns have 7000

(Nurmes) and 10 000 (Lieksa) inhabitants and are surrounded by large sparsely populated

rural areas. The proportion of young people is low compared to Kuopio. The share of young

people (aged 15-24) in the population of both towns is around 10%. In Nurmes and Lieksa

the population has relatively low education and income, and the unemployment level is

MYPLACE August 2014


twice as high (or even more) as the average in the country. The towns are located in a

declining socio-economic area with decreasing economic activity and ageing population. The

future prospects of young people living in this area are limited because of a lack of higher

and even vocational education and employment opportunities.

Both towns are relatively homogeneous in terms of ethnicity of their population. However,

Lieksa has received some hundreds of asylum seekers, who have gotten residence permits

during the last four years. Immigrants (mainly from Somalia) have caused suspicion and

conflicts in this relatively small and closed community, which has a history of ethnic tensions

starting from the late 1950s (expulsion of the Roma in 1956). In Nurmes and Lieksa, the level

of support for The Finns is relatively high at nearly 25% of the voters, which is above the

national average (19 percent).

Georgia

The survey locations in Georgia were Kutaisi (more specifically, Kakhianuri district of the

city), situated in the western part of the country, and the town of Telavi, situated in the

east. The distance between Kutaisi and Telavi is approximately 300 km.

Kutaisi is the capital of the Imereti region, the second largest city in Georgia (with 194.7

thousand inhabitants according to the latest data of the National Statistics Office of

Georgia), and one of the biggest industrial cities of the country. The National Parliament of

Georgia moved to Kutaisi in October 2012, when the MYPLACE survey fieldwork was

conducted. It was expected that this would lead to important changes in the life of the city,

since more attention would have been paid to its infrastructure, more social and cultural

events would have been organized, and potentially, there would have been more job

opportunities as well. The ‘supply side’ of all types of political movements, including, of

course, the radical ones, was also expected to significantly increase, and try to influence the

local population (although it was already rather high in Kutaisi even before the Parliament

moved there).

Telavi, on the other hand, is the capital of the Kakheti region and a much smaller town with

a population of less than 30 thousand inhabitants. Traditionally, political life is significantly

less active in Telavi (compared to the situation in Kutaisi); hence, the ‘supply’ factor in

radical politics is less prominent. Other important differences are observed from the point

of view of infrastructure and living standards. Employment opportunities are more limited,

as are the sources of information about current events (including local sources of

information).

The eastern and western parts of Georgia are traditionally characterized by different styles

and rhythms of life, as well as by differences in the dominant attitudes of the population

towards major events and/or social and political institutions. The population of Imereti is, in

MYPLACE August 2014


general, believed to be rather active, while the population of Kakheti is considered to be

more passive. Both desk research and expert interviews conducted in terms of WP1 and

WP3 suggested the same. Hence, different level of engagement of young people were

expected to be documented in these two locations.

Eastern Germany

In eastern Germany, the WP4 survey was conducted in Jena in Thuringia and Rostock in

Mecklenburg-Vorpommern (MV), both parts of the former German Democratic Republic.

Jena is the second largest city in Thuringia with slightly over 100 thousand inhabitants

(105,463). It is comprised of a large number of young people and students. The historic

Friedrich-Schiller University and the Fachhochschule host together about 25 thousand

students. It is therefore not surprising that in 2010, the age group 15-25 made up 13.4

percent of the city’s population (Statistische Ämter des Bundes und der Länder, 2012). Over

the last decade, the foreign population in Jena has grown to a share of 6.5 per cent in 2011,

which is almost twice as high as in Rostock (Thüringer Landesamt für Statistik).

After reunification of Germany, Jena developed into a prosperous site for specialised

industries, including the optical, chemical, pharmaceutical and medical industries

(Wirtschaftsförderungsgesellschaft Jena mbH, 2013). The unemployment rate is lower than

in the rest of Thuringia, in eastern Germany, and even in western Germany.

Jena has a reputation for being a heartland for right-wing extremism. The Festival of the

People, the largest right-wing extremist event in Europe, took place in Jena between 2005

and 2007. Core members of the right-wing terror cell Nationalsozialistischer Undergrund

(NSU) originally came from Jena and were part of the Sektion Jena of the Thüringer

Heimatschutz, a radical right-wing group, during the 1990s (Kynast, 2005; Zander, 2005).

With slightly over 200 thousand inhabitants (202,734), Rostock is the only major city in MV,

which has a predominantly rural character. Whilst the countryside in MV is acutely affected

by population loss, an ageing society and a low share of the foreign population, Rostock was

able to reverse shrinkage and experience moderate population growth (Scholz et al., 2010).

Because of its educational offerings, including a historic university with 15,000 students,

research institutes and related businesses, Rostock has been able to attract a young

population. In 2010, Rostock comprised a higher proportion of young people (12.4 per cent)

than the German average (11.1 per cent) (Statistische Ämter des Bundes und der Länder,

2012).

MYPLACE August 2014


In contrast to Jena, Rostock has faced economic decline since the German reunification. Like

Jena, Rostock has been stigmatised as a stronghold for right-wing extremism. Twenty years

ago, in the summer of 1992, a group of 100 people attacked the Zentrale Aufnahmestelle für

Asylbewerber (ZAST), an asylum for Vietnamese guest workers in Rostock-Lichtenhagen

(Prenzel, 2012).

Western Germany

In western Germany, the WP4 survey was conducted in Bremen wards and Bremerhaven

wards. The selection of locations was driven by the basic interest of the MYPLACE project to

learn more about young people’s involvement in radical, extreme or populist movements.

Drawing from the WP1 summary and German reports, as well as the knowledge of local

factors potentially not captured in the broader literature, it has been concluded that the

two most important criteria for being influenced by radical movements in Germany were:

(1) Community segregation/integration and (2) Socio-economic inequalities (‘deprivation’).

Based on this, the western German team agreed on the following composition of the

localities:

1) Bremen wards: Alte Neustadt, Altstadt, Arbergen, Arsten, Aumund-Hammersbeck,

Barkhof, Borgfeld, Buntentor, Bürgerpark, Burg-Grambke, Ellener Feld, Ellenerbrok-

Schevemoor, Farge, Fesenfeld, Findorff-Bürgerweide, Gartenstadt Süd, Gartenstadt Vahr,

Gete, Grolland, Habenhausen, Hastedt, Horn, Huckelriede, Hulsberg, Kattenesch, Lehe,

Lehersterdeich, Lesum, Mahndorf, Neuenland, Neu-Schwachhausen, Neustadt,

Oberneuland, Osterholz, Ostertor, Peterswerder, Rablinghausen, Radio Bremen,

Regensburger Str., Rekum, Riensberg, Rönnebeck, Schönebeck, Schwachhausen,

Sebaldsbrück, Seehausen, St. Magnus, Steintor, Südvorstadt, Vegesack, Weidedamm.

2) Bremerhaven wards: Königsheide, Fehrmoor, Leherheide-West, Speckenbüttel,

Eckernfeld, Twischkamp, Goethestraße, Klushof, Schierholz, Buschkämpen, Mitte-Süd,

Mitte-Nord, Geestemünde-Nord, Geestendorf, Geestemünde-Süd, Bürgerpark, Grünhöfe,

Schiffdorferdamm, Surheide, Dreibergen, Jedutenberg.

Greece

In Greece, the WP4 survey was conducted in two suburbs of Athens, Municipal Departments

of Argyroupoli and Nea Philadelphia. They were selected based on their population size

(around 30,000, i.e. realistic to cover during the fieldwork and, at the same time, large

enough to conduct the expected number of interviews) and their characteristics regarding

MYPLACE August 2014


community segregation/cohesion, socio-economic inequality, “supply” of radicalism,

political heritage and civic engagement.

Argyroupoli is situated in the relatively affluent south-eastern region of Athens and is

populated by representatives of middle and upper-middle classes. Although there are no

official data available regarding socio-economic inequality within and between the selected

areas, the existing information suggests higher income and standard of living in Argyroupoli

where real estate is significantly more expensive than in Philadelphia.

Argyroupoli has significant development potential due to the new metro station, which

opened to the public in the summer of 2013, and the neighbouring area of the former US

military air base, which is a public asset and is considered to be the largest urban waterfront

property in Europe. Politically, population of Argyroupoli is characterized by centrist-liberal

tradition.

Nea Philadelphia, on the other hand, is located in the poorer north-western part of the city

and has characteristics of a lower-middle and working class area. It includes

neighbourhoods that are vulnerable to social deprivation and exclusion such as a social-

housing district and parts of the refugees’ old neighbourhoods. The latter were populated

by Greek refugees who had been expelled from the Asia Minor coastline after the First

World War, following the Greek-Turkish war of 1919-1922, and the “Catastrophe of Smyrna”

in 1922. Philadelphia has a left-wing political tradition due to the working class origin of its

population.

According to 2001 census data, the vast majority of the population in both locations is of

Greek origin. Both locations exhibit similar voting patterns in national and local elections

which also follow the general voting pattern in Greece. However, in the context of the

current crisis, it is anticipated that radicalism will be present in both locations. In both

locations there are open assemblies and the residents have mobilized against austerity

measures. In Philadelphia, however, such activism seems to be more intense and there is an

anarchist occupation. In both locations there are citizens’ associations and clubs, while

during the crisis there has been an increase in participation in voluntary initiatives.

Hungary

In Hungary, the WP4 survey was conducted in Sopron and Ózd. Both towns are on the

border and far from the political, economic, and cultural centre, Budapest. However they

differ from each other in a number of respects.

MYPLACE August 2014


Sopron is located on the Hungarian-Austrian border and has 61,072 inhabitants (Hungarian

Central Statistical Office, 2011). Since Austria is a developed country, its vicinity offers more

advantages in trans-border relations (tourism, employment, trade). Sopron has historic civic

traditions. There are several secondary schools and university departments in Sopron. The

economic indicators of Győr-Moson-Sopron County are above the national average. The

average unemployment rate during 2011 was 4.6%. The average annual unemployment rate

is the third lowest in Hungary, just behind the capital and Pest County. The proportion of

Roma population in Győr-Moson-Sopron County was 2.7% in 2003.

Ózd is located near the Hungarian-Slovakian border and has 34,072 inhabitants (Hungarian

Central Statistical Office, 2011). Historic civic traditions are absent; it is a typical socialist

industrial town. There are only two secondary schools in Ózd. It has the third largest Roma

community in Hungary according to 2001 census data (according to some estimates, 37% of

the population is Roma). Unemployment, which has affected many since the fall of socialism

in this industrial and mining region, doubled between 2002 and 2009. A significant

proportion of the population is not highly educated. Earlier they worked as semi-skilled

workers, and they are currently long-term unemployed.

Latvia

In Latvia, the WP4 survey was conducted in urban areas of Riga and Daugavpils. Āgenskalns

apkaime was the area surveyed in Riga. In Daugavpils city, Forštate (Jaunā Forštate and Vecā

Forštate) and a part of the Jaunbūve district were used. The selected areas present

contrasting cases in terms of economic development, ethnic composition, and social

situation.

Āgenskalns is one of the 58 "apkaimes" (localities) in Riga, the capital of Latvia, and the most

economically developed region in the country. There are more jobs available, salaries are

significantly higher, and the overall economic situation is arguably the best in Latvia. There

are lots of opportunities for young people to get involved in different social, civic and

political activities. Riga, and Āgenskalns apkaime in particular (due to several university

branches located there), has quite diverse population, since it hosts a mix of people from

different regions that come to work or study there. Thus, people in Āgenskalns seem to be

more open, liberal, and tolerant to different ideas and cultural influences. The ethnic

proportions of Russians and Latvians are similar, yet Āgenskalns is seen as one of the

predominantly Latvian parts of the city. Āgenskalns has a number of schools and

kindergartens, a university branch, one of Riga’s largest hospitals, several churches and

museums, an old and lively market, one of the city’s largest parks; it is very well connected

to other districts and to the centre of the city. The number of inhabitants in Āgenskalns

district is 28,741 (Riga city municipality data, 2010).

MYPLACE August 2014


Daugavpils is located in the Latgale region. For a long time, this region was considered as

lagging behind others in terms of economic development, infrastructure, employment,

innovation, etc. The emigration rates from this region are very high, especially among the

youth. Latvians are quantitatively one of the smallest groups in this city, while 85% of the

population are ethnic minorities. The ethnic composition of the population and the

closeness of the city to Russia, in combination with economic struggles of the population,

provides fertile ground for the rise of radical ideas.

The areas selected in Daugavpils city are similar to each other, yet have their distinct history

and character. Jaunā Forštate, like the city itself, has mostly Russian population, and a

comparatively high number of young people live there. Transport connections to other parts

of the city have been relatively poor until recently, but the infrastructure is improving. There

are kindergartens, a school, a few churches (Russian Orthodox and Catholic), as well as a

cultural centre in this district.

Vecā Forštate is a small area neighbouring Jaunā Forštate. Due to its small size, in some

sources it is not even distinguished from the latter, but merged with Jaunā Forštate as

“Forštate”. Historically, Vecā Forštate is one of the oldest districts of Daugavpils, and the

19th century ancient wooden architecture has been partially preserved here. There are

significant numbers of Old Believers and pensioners among the locals, and very few young

people. Almost all buildings are private homes, and gardening is a popular occupation. The

local school was recently closed, yet the “Cultural Centre” is still functioning.

Jaunbūve began to develop in the second half of the 19th century thanks to the development

of the nearby railroad. Many and varied economic, social and spiritual activities are

concentrated there. A symbol of multiculturalism and mutual respect between different

groups of the local society is the “Church hill” where four churches of different

denominations stand right next to each other. Jaunbūve district is one of the largest districts

in Daugavpils, thus only a part of it was selected for the MYPLACE survey – the one

bordering Jaunā Forštate. The total population of the merged Forštate/Jaunbūve area was

28,680 inhabitants (Daugavpils municipality, 2011).

Portugal

In Portugal, the WP4 survey was conducted in Telheiras and Barreiro. The first site consists

of a recently built neighbourhood located in the north of Lisbon. Telheiras is clearly a

“middle class” neighbourhood, with professional population, with corresponding lifestyle

and consumption patterns. Despite being recently established, this neighbourhood has a

MYPLACE August 2014


relatively vibrant community life, stimulated by local institutions and associations. Electoral

behaviour is quite similar to the national average.

The second location, Barreiro, is a small city located on the south shore of the river Tagus

that has been undergoing a period of economic decline during the last few decades.

Barreiro was the location of a large industrial conglomerate, CUF – with 10,000 workers

living in factory villages in the 1960s (Almeida, 1993: 16). Barreiro was also the headquarters

of the Portuguese Railways maintenance workshops, and was called “Portuguese Barcelona”

during the dictatorship (ibid: 17). Barreiro’s population is positioned more towards the left

wing of the Portuguese political spectrum than that of Telheiras.

Due to sampling constraints, the two locations had to be enlarged, while maintaining

basically the same contrasting characteristics. The Telheiras location was enlarged to include

Lumiar, the civil parish where it is integrated, which is inhabited by 41,163 residents

(National Population Census 2011). Its population is relatively young (26% below the age of

25 years old, with 11% of these people aged between 15 and 24 years old. Barreiro, was

initially comprised of two civil parishes and then was enlarged to four (Barreiro, Lavradio,

Alto do Seixalinho e Verderena). These four parishes contain 52,157 inhabitants. Population

of this location is slightly older, with 22% being under the age of 25 years old and 9% (4,767

people) of its residents aged between 15 and 24 years old. The educational profiles of the

two populations are strikingly different: in Telheiras and Lumiar, 42% (17,218) have a

tertiary degree in contrast to Barreiro, where only 11% (5,567) have such qualification.

Portugal’s average is 17%.

Russia

In Russia, the WP4 survey was conducted in: 1) Kupchino – one of six municipal districts

included in the Frunzensky administrative district of St. Petersburg, and 2) Vyborg – a

municipal district which is located in the Leningrad oblast. The first is located within a city of

St. Petersburg, which is the centre of the north-western region of Russia, and the second

most important city in the country after Moscow, and the second is a smaller town in the

region bordering Finland.

Both sites are interesting historically, although they have different historical backgrounds

and represent different relationships to history – importance in Kupchino, ambivalence in

Vyborg. In both cases, migration issues and interethnic tension are important issues.

Kupchino is the unofficial name for the main area falling within the Frunzensky

administrative district of St. Petersburg. It is a typical suburb (“sleeping district”) with huge

Soviet housing blocks constructed in the 1960s-80s. The selected area – the municipal

MYPLACE August 2014


district with the name “Kupchino” (#73) is inhabited by 54,500 people, with 6,300 young

people among them. It is located in the centre of Kupchino. It experiences significant

problems with public transportation – the Kupchino district has only two underground

stations on the edges of its territory for 400,000 people living in the Frunzensky district. It is

physically separated from the rest of the city by two railways in the east and the west, as

well as an industrial “grey zone” in the north, separating the district from the city centre.

While Kupchino may not be objectively the most socially deprived district of St. Petersburg,

it suffers relatively significant deprivation because of the isolation caused by transportation

problems and the urban mythology surrounding the district (it is perceived as criminal and

marginalized). The neighbourhood is ethnically mixed and experiences a significant impact

due to immigration from the southern ex-Soviet republics who tend to concentrate here

because of relatively low apartment prices.

The second location is the town of Vyborg in the Leningrad oblast, which is a municipality

with 80,700 inhabitants (number of young people – 11 400). Vyborg is interesting because

of its specific history and heritage, on the one hand, and its economic and political situation,

on the other. The town became part of the Soviet Union in 1944. This medieval town and its

castle were built by the Swedes; it subsequently became the second largest town in Finland.

After 1944, the Finns were expelled and their homes occupied by new inhabitants. Today,

history is a complex issue here – the locals appear alienated from the medieval history of

the town, although a number of historical reconstruction clubs actively try to engage young

people in Vyborg.

Slovakia

In Slovakia, the WP4 survey was conducted in the town of Trnava and a district of Rimavská

Sobota. This selection of these locations was based on the assumption that the most

important factors for the growth of political extremism and radicalism are social and

economic deprivations of the regions and communities on the one hand, and the possibility

of potential ethnic conflicts on the other hand. The selected locations differ significantly in

these regards.

Trnava has a homogenous ethnic structure and a relatively low level of social and economic

deprivation. The location ranks among “strongly developing” areas with regard to its

economic situation. Factors conditioning this include its proximity to the capital and the PSA

Peugeot automobile factory. Several universities are located here, therefore, the

concentration of youth population is high.

On the other hand, the selected district of Rimavská Sobota is characterized by high ethnic

heterogeneity as well as by above-the-average level of social and economic deprivation.

MYPLACE August 2014


Several ethnic groups meet in this location – Slovaks, Hungarians, and Romany. The area

ranks among the poorest locations in the country (high unemployment and poverty rates,

social issues). Problems have accumulated there for several decades, and are mainly caused

by lack of economic opportunities.

Spain

In Spain, the WP4 survey was conducted in two locations in the autonomous community of

Catalonia: Vic and Sant Cugat. These locations have been selected based on the following

criteria:

(1) Presence and electoral success of Plataforma x Catalunya (a relatively new,

populist, right-wing and xenophobic political party. Its electoral support, for now,

is based in Catalonia);

(2) Socioeconomic deprivation and the impact of the current economic crisis;

(3) High immigration rates and the history of previous ethnic conflicts; and

(4) Presence of “The Indignant” (or 15-M) movement through local assemblies and

commissions.

Local 15-M (“The Indignant”) movement exists in both locations. The city councils have

developed active youth policies, and the main party is CiU (the center-right Catalan

nationalist political party, now governing Catalonia as well).

Vic has a population of about 40,000. In 2010, 25% of the inhabitants of Vic were

immigrants, compared to Catalonia’s overall 16%. A significant proportion of immigrants are

from Morocco and Sub-Saharan Africa. According to the labour union UGT, the foreign-born

citizens of Vic suffered a 43% unemployment rate, as compared to the average rate of

17.8% for Catalonia. Vic is the seat of a small university (Universitat de Vic) and has a

dynamic civil society.

Vic is governed by a coalition of CiU (its electoral support is 32%) and ERC (a center-left

secessionist party with 20% of the votes). Two more political parties are active in Vic and

demonstrate the radicalization in politics on both ends of the left-right spectrum: PxC and

the CUP. Each of these political parties has different party manifestos, action strategies and

discourses about immigrants. On the one hand, Plataforma per Catalunya (PxC) is a populist

right-wing party generally considered by academics, the mass media and public opinion as

extremist and xenophobic. PxC holds a firm stance on the control of immigration. Its leader,

Josep Anglada, was born in Vic, and the party has 5 seats in the town council of Vic, which is

the largest representation the party has in any local government.

At the opposite extreme, the Candidatura d'Unitat Popular (CUP), a leftist secessionist party

that promotes a more participatory democracy, has two representatives in the town

MYPLACE August 2014


council. Young people are over-represented in the CUP, enjoying a horizontal organisational

structure. CUP is not famous for anti-immigration discourse, but the very fact that it defends

self-determination of Catalonia might represent a basis to consider it as not being

sufficiently inclusive, and ignoring certain sectors of the population. In sum, Vic has a

greater propensity to be radicalized than other places in Catalonia.

Sant Cugat is an extremely affluent town to the north of Barcelona, with a high level of

investment from high-tech companies. Its central district, the most densely populated, has a

population of 56,000. Sant Cugat has a registered population of immigrants of only 14%, half

of which are Latinos, which is not characteristic for Catalonia. For linguistic and cultural

reasons, immigrants from Latin America integrate more easily than foreigners from other

countries outside Europe. In other words, there are no a priori reasons to suspect a high

propensity for radicalization in Sant Cugat. In the last regional elections, the electoral

support for PxC was only 0.58% (5.82% in Vic).The center-right Catalan nationalist party CiU

governs Sant Cugat with a majority in the town council.

UK

The two locations for the WP4 survey in the UK were Coventry and Nuneaton. These

locations are geographically close to one another, only 13 km apart, in a wider area

commonly known as The West Midlands.

In Coventry, parts of Foleshill and St. Michaels wards were surveyed. These areas extend

north-east from the city centre. This location represents an inner city with high density

housing and a variety of housing types. The majority of the dwellings consists of Victorian

terraced homes with a number of high rise flats. Foleshill and St. Michaels have large

student populations (who attend either Coventry University, part of which is within the

research location, or Warwick University) and a number of residence halls. These were

excluded from the sample. Both wards are ethnically mixed and are characterized by low

levels of radical political activity.

This location consists of 13 Lower Super Output Areas (LSOA), which are the smallest

administrative units geographically available. The total population for these areas was

25,451 (Census 2011), with 14,404 aged 15-24 years old. There were 9,092 households

within the research location.

The wards of Arbury, Camp Hill, Kingswood, Abbey, and part of Wem Brook were selected

for fieldwork in Nuneaton. These wards cover approximately half of Nuneaton in the north

and west of the town. This location is a Market Town, with mixed housing and is

predominately white British.

MYPLACE August 2014


The research location consists of 26 Lower Super Output Areas (LSOA). The total population

for these areas was 40,612 (Census 2011), with 7,064 aged 15-24 years old. There were

17,360 households within the research location. In a number of selected wards there has

already been significant radical right wing activity including the election of local councillors

for the Nuneaton and Bedworth Councils. Levels of deprivation varied across the research

location, with a number of LSOAs within the 10% most deprived and least deprived

nationally.

3.3 Ethics

While collecting data for the work packages 4 and 5, all partners were required to follow the

confidentiality procedures in line with laws on data protection.

Partners provided background information by distributing/sending information

letters/leaflets/brochures to the potential respondents prior to visiting their homes. This

way respondents could get information about the MYPLACE project in advance of fieldwork.

These letters often enclosed a telephone number, allowing time and opportunity for

potential respondents to seek further information, to ask additional questions or to refuse

an interview.

Before the interviews, all survey respondents were informed how the collected data will be

processes, stored and used and how their identity will be protected. Furthermore, they

were informed that they can withdraw from participating at any time or refuse to answer

any question they do not want to answer. It was made clear to all respondents that

participation in the survey was voluntary.

All respondents provided informed ‘verbal’ consent, after the researchers provided full

information about the project and made sure that the respondents understand that the

data they provide will be kept confidential and the names and addresses will be separated

from substantive data in both its computer and paper records. In addition, In order to

ensure data confidentiality, all field researchers/interviewers signed a confidentiality

statement. With their signature, each interviewer confirmed the protection of anonymity of

collected information.

Furthermore, each partner selected 60 respondents from the WP4 survey in order to

conduct further semi-structured interviews for WP5. As part of the survey, the interviewers

collected information on young people agreeing or refusing to take part in the follow-up

qualitative interview, by filling in specifically prepared contact forms. Nevertheless, it has

been ensured that identification of a specific individual in a dataset is not possible, as all

personal information has been removed from the dataset and each respondent was

MYPLACE August 2014


assigned a unique ID code. The sample of the respondents for the qualitative semi-

structured interviews was drawn from a list of those respondents who accepted to

participate and the qualitative interviews were carried out by other team of researchers.

Finally, some partners used incentives to encourage the survey participants to further

participate in the semi-structured interviews.

3.4 Sampling

Three basic sampling plans were developed for the partners, and each partner used the plan

best suited for their situation. In survey locations where reliable lists of individuals aged 16-

25 years old and their contact information were available, samples were drawn from these

lists (Sampling Plan 1). In survey locations where such lists were not available, but reliable

lists of households were accessible (for example, from post code address files, or electrical

grids), samples of households were drawn. When contact was made with a household

containing more than one member of the target age group, the respondent was selected

using the last (or next, in case of Latvia) birthday method (Sampling Plan 2). In survey

locations where neither reliable lists of individuals nor reliable lists of households were

available, households were selected using random route sampling, and individual household

members were selected using the last (or next, in case of Latvia) birthday method (Sampling

Plan 3).

Sampling from lists of individuals and Sampling from lists of households: The lists of

individuals, as well as the lists of households and the information contained in each varied

between countries. Most of the areas selected for the survey consisted of multiple smaller

units, such as administrative districts or voting precincts; some even consisted of multiple

small settlements. Where the lists of individuals included the subunit of residence, the

overall target sample size was allocated among subunits proportional to population size,

and samples were drawn separately from each unit. For those lists that did not contain

information about geographic subunits, a simple random sample of individuals or

households was made.

There was a wide variance in expected response rates across the survey locations; thus,

some partners have tried to minimize non-response by sending letters informing the target

respondent or his/her household about the survey prior to attempted contact; advertising

the survey in the local media; or offering small incentives.

In instances of sampling from lists of households, the households where no member of the

target population was present and, consequently, where no interviews could have been

conducted, were declared as ineligible households.

MYPLACE August 2014


Random route sampling of households: In cases where no lists of individuals or households

were available, households were selected using random route sampling, a method of

systematic sampling which approximates a simple random sample. Starting points of the

interviewers’ routes were determined. From these starting points, a prescribed walking

pattern was followed, with a given step size and starting household number. For example, if

the starting household number was three, and the step size was five, then the interview

started the route in the 3rd household on his/her route, and then every 5th household on

his/her route was targeted for interview.

Random walk methodology developed by CRRC-Georgia (and presented in D4.2 report) was

provided to all partners that were using Sampling Plan 3, but the walking patterns varied

between the survey locations, because some partners and/or their sub-contractors

conducted fieldwork applying their own protocols. In all such cases, the proposed random

walk protocols were reviewed by the work package coordinators, and, if necessary were

modified for the MYPLACE survey.

The sampling strategies used in each survey location are presented in Table 3.4.1.

Table 3.4.1. Sampling strategies by MYPLACE survey locations5

Country Survey location Sampling plan

Croatia Podsljeme (CRO) 1

Pescenica (CRO) 1

Denmark Odense East (DEN) 1

Odense Center (DEN) 1

Estonia Narva area (EST) 1

Tartu (EST) 1 and 3

Finland Lieksa and Nurmes (FIN) 1

Kuopio (FIN) 1

Georgia Kutaisi (GEO) 3

Telavi (GEO) 3

Western Germany Bremen (GER-W) 1

Bremerhaven (GER-W) 1

Eastern Germany Jena (GER-E) 1

Rostock (GER-E) 1

Greece New Philadelphia (GRE) 3

Argyroupouli (GRE) 3

5 In case of Sampling plan 1, the samples were drawn from the lists of individuals aged 16-25 years old. In case

of Sampling plan 2, the households were sampled based on the lists of households, and in households with more than one member of the target age group, the respondent was selected using the last (or next, in case of Latvia) birthday method. Random route sampling was employed in survey locations where neither reliable lists of individuals nor reliable lists of households were available (Sampling Plan 3).

MYPLACE August 2014


Hungary Sopron (HUN) 1

Ozd (HUN) 1

Latvia Agenskalns (LAV) 3

Forstate and Jaunbuve (LAV) 3

Portugal Lumiar (POR) 3

Barreiro (POR) 3

Russia Kupchino (RUS) 2

Vyborg (RUS) 2

Slovakia Rimavska Sobota (SLO) 3

Trnava (SLO) 3

Spain Vic (SPA) 1

Sant Cugat (SPA) 1

United Kingdom Coventry (UK) 2

Nuneaton (UK) 2

For all three sampling plans, in most cases, initial contact attempts were made by visiting

the households. When sampling from lists of individuals, sometimes initial contact attempts

were made by telephone. This was only done to arrange a meeting time and place for a

face-to-face interview. In order to minimize non-response, a minimum of three contact

attempts were made by all partners. Contact attempts subsequent to the initial one could

have been made by telephone, but only to arrange a time and place for a face-to-face

interview.

In addition to recording non-response rates, the partners following Sampling Plan 2 and

Sampling Plan 3 had to estimate the proportion of households that did not include a

member of the target population, because these were considered as ineligible households

during the fieldwork.

Each partner was expected to achieve 600 completed interviews per location. Several

partners made some variations of the original sampling approach during the fieldwork,

mainly with the aim to achieve the expected number of completed interviews.

In Croatia, changes were made during fieldwork in Peščenica as the contact rate was lower

than expected. The number of sample units was increased in this location by expanding the

geographic area under consideration. In other words, five adjacent local committees were

added to the original location, which increased the number of sample units from 15,000 to

22,000 (in order to obtain 600 completed interviews).

In Finland, additional samples were drawn in both locations, since the interviewers were

facing difficulties in reaching the potential respondents, and there was a high refusal rate.

However, even after the additional sample was employed, the targeted number of

MYPLACE August 2014


interviews was not achieved, and finally the whole population of the target age group was

included in the sampling frame.

Some partners had to expand the areas of the selected locations because of low response

rates. For example, in Bremenhaven in western Germany, additional districts had to be

selected during the fieldwork in order to conduct more interviews. In Portugal, which used

the random walk sampling strategy, an extremely high rate of household ineligibility was

documented (no household members of the targeted age), as well as high refusal rates.

Both issues led to the decision to expand the initial borders of the field sites. In other cases,

partners had to reduce the original step size. The Greek team, which also used the random

walk sampling strategy, had to change the original sampling methodology by reducing the

step size, and interviewed all eligible persons in a household instead of only one member

per household.

In Latvia, every residential address in the selected locations was included in the sampling

frame. That is, interviewers searched for a young person of the targeted age in every

household in the areas of Āgenskalna apkaime in Riga and Forštate and Jaunbūve district in

Daugavpils. Accordingly, the survey covers all households in these areas, without exception.

In instances where there was more than one person aged 16-25 in the household, the

selection of the respondent was based on the nearest birthday method.

Two different sampling plans were used in Estonia. In Narva, Sillamäe and Kohtla-Järve,

random sampling based on the national population register was used. In Tartu, random

sampling was used at the beginning of the fieldwork. After 262 interviews were conducted

based on this sampling approach, the survey team switched to random route sampling and

the rest of the respondents were located using the latter method. To ensure 600 interviews,

parishes near Tartu were also included in the sampling frame.

Table 3.4.2 summarizes information about MYPLACE survey fieldwork in all locations.

Table 3.4.2. MYPLACE survey fieldwork

Country Survey locations N of

interviews Fieldwork dates

MYPLACE August 2014


from to

Croatia

Podsljeme 610 September 14, 2012 December 20, 2012

Peščenica 606 September 15, 2012 December 18, 2012

Denmark

Odense Centre 402 October 14, 2012 April 10, 2013

Odense East 380 October 30, 2012 April 10, 2013

Estonia

Narva area 617 October 28, 2012 January 21, 2013

Tartu 634 October 28, 2012 February 26, 2013

Finland

Lieksa and Nurmes 452 September 11, 2012 February 26, 2013

Kuopio 430 September 13, 2012 March 9, 2013

Georgia

Kutaisi 579 October 13, 2012 November 18, 2012

Telavi 588 October 10, 2012 November 15, 2012

Western

Germany

Bremen 604 September 24, 2012 January 21, 2013

Bremerhaven 332 October 8, 2012 March 12, 2013

Eastern Germany

Jena 608 January 22, 2013 March 30, 2013

Rostock 608 February 15, 2013 May 3, 2013

Greece

New Philadelphia 600 February 25, 2013 March 27, 2013

Argyroupoli 595 February 28, 2013 April 15, 2013

Hungary

Sopron 597 October 3, 2012 December 10, 2012

Ózd 590 October 1, 2012 December 16, 2012

Latvia

Agenskalns 600 October, 13, 2012 March 12, 2013

Forstate and Jaunbuve 600 October, 13, 2012 March 29, 2013

Portugal

Lumiar 596 September 13, 2012 December 11, 2012

Barreiro 594 September 14, 2012 December 11, 2012

Russia

Kupchino 599 September 27, 2012 October 29, 2012

Vyborg 600 September 29, 2012 October 24, 2012

Slovakia

Trnava 600 September 18, 2012 November 16, 2012

Rimavská Sobota 600 September 20, 2012 November 11, 2012

Spain

Vic 597 November 7, 2012 December 28, 2012

Sant Cugat 592 November 7, 2012 December 28, 2012

United Kingdom Coventry 542 October 8, 2012 December 19, 2012

Nuneaton 550 September 26, 2012 March 12, 2013

In research locations where some form of address based sample, as opposed to a list based

system, was used, researchers were faced with the choice of interviewing all or one eligible

household members. Each strategy reduces sample precision. The former strategy, which

involves interviewing all eligible household members, is arguably subject to an intra

household correlation as there is a tendency for there to be similarities within rather than

between households. The latter strategy, where one respondent per household is used,

avoids the intra household correlation problem but precision is lost through excluding

eligible respondents (ie. variation in the selection probabilities). Each household sample is

therefore subject to some loss of precision and a consequential reduction in the effective

sample size.

MYPLACE August 2014


3.5 Weighting

Most partners decided not to apply any weighting to their data, either because of the lack of

reliable local demographic data, or due to full coverage of the households in the survey

locations. The Croatian, Greek, Russian and Georgian teams could not obtain reliable

statistical data, and hence, chose not to apply any weighting to their data. The eastern

German team considered data weighting with regard to educational status and age,

however, due to the lack of adequate reference statistics weighting was not applied. The

western German team intended to compare the social distribution of Bremen and

Bremerhaven datasets to the 2011 census conducted in the selected wards. In case of big

variations from the census, they would have calculated weights to adjust representation of

specific demographic groups. However, since the Statistical Office of Bremen was not able

to deliver the necessary data for comparison, partners from western Germany conducted

data analysis without weighting.

Partners from Denmark also chose not to weight their data. They estimated that the

distribution of male and female respondents was fairly representative. However, there was

an issue of over-representation of well-educated young people, and under-representation

of young people with no education, and of young people representing ethnic and religious

minorities. Because no accurate information was available about the level of education by

age groups, and about the different ethnic groups, it was not possible to weight the data.

In Latvia, demographic data were only available at the city level, so there was no reliable

information available on the demographic characteristics of youth in the selected areas of

the cities. Taking into consideration the applied sampling strategy, the only design effect in

the case of the sample arose from the clustering of respondents within households, which

resulted in differing selection probabilities (i.e., young people living in households where

there was more than one young person aged 16-25 had a lower probability to be

interviewed). Although these probabilities could have been calculated, and adjusted for by

applying weights, unfortunately, no data on the composition of the households were

collected during the survey; therefore, the data were not weighted. Comparison of the

sample characteristics with demographic data provided by the Central Statistical Bureau of

Latvia on Riga and Daugavpils allowed for the conclusion that the age and gender

composition of the samples was similar to that observed in the respective cities. Thus, the

fact that no weights were applied in the WP4 data in Latvia should not compromise

representativeness of the results, or cause a notable bias in the distribution of answers.

MYPLACE August 2014


In Spain, weighting was not needed because socio-demographic characteristics of the

sample were quite similar to the entire 16 to 25 year old population in both locations.

Similarly, in the UK, an analysis was performed to make an informed decision about whether

weighting was appropriate for the UK set. Data from the 2011 census (at LSOA) were used

to compare both gender and ethnic distributions with the sample characteristics. Because

distributions were similar, it was decided that no demographic weighting of the data would

be undertaken.

Estonia, Slovakia and Hungary also have not weighted their datasets. Only two partners –

Finland and Portugal – weighted their data sets.

In Finland, the achieved sample was biased in regard to the gender and age of the

respondents. The gender bias was so moderate that it was not used as a basis for weighting;

this bias was also in line with most social science surveys, and attributed to the greater

reluctance of men to participate. The data were weighted to adjust the age distribution of

the respondents, since the youngest subgroup (16-17 year olds) was significantly

overrepresented in both Kuopio and Pielinen Karelia samples (see Table 3.3).

Finally, in Portugal the samples in both survey locations were somewhat skewed in terms of

age and gender. Both factors led to a significant risk of erroneous and misleading

interpretations of the data. Lumiar’s sample was composed of individuals much older than

expected and with a large gender imbalance (more women than men). In contrast, the

Barreiro sample was much younger than expected and also unbalanced in respect to gender

(more men than women). The execution of a one-sample Chi-square adherence test

(p<.0001) for each location. Led to the conclusion that the survey sample had significant

deviations and did not represent the target population.

To solve this problem, three different weights were created, one weight for the entire

sample and two – for each location.

3.6 Data processing and cleaning

Data entry

Some partners used PAPI during the fieldwork, and entered the data after fieldwork.

Students experienced in data entry were employed by the Croatian team and their

collaborator, the Institute of Social Sciences Ivo Pilar. In Finland, an initial quality check of

the completed questionnaires was performed during the data entry process. In some cases,

self-administered questionnaires were of poor quality (e.g. respondents had given a lot of

“don’t know” answers or had selected code “3” in every box in the questionnaire). The

poorest questionnaires were disqualified.

MYPLACE August 2014


CRRC-Georgia used CSPro software for data entry. In order to ensure the highest quality of the data, CRRC employed double blind data entry. After the comparison of the two data files, the CSPro data file was transferred into SPSS. The Greek team used CONVERSO software for data entry. In the case of CAPI interviews, the data entry stage was skipped. In the UK data, for example, in the majority of cases, data were captured using the CAPI MS Access Database. Data files were transferred by the fieldworker on a daily basis via a data upload to dropbox. The supervisor then checked for completeness of the files, and the national coordinator transferred the questionnaires to a master file, which was used for monitoring purposes. In the few cases (50) that the data were captured via a paper questionnaire, forms were collected and sent to MMU for data entry directly into the master CAPI file by an administrator. These were then randomly checked to ensure data quality.

Data cleaning

In cases of both PAPI and CAPI, data cleaning was performed by the partners. Data cleaning

protocol and a corresponding syntax file were developed by the MMU team which provided

guidance on how to deal with missing cases/answers, ensure logical consistency of each

respondent’s answers, etc. Incomplete questionnaires with a high number of missing

answers were to be deleted from the datasets according to the MMU protocol, along with

the questionnaires completed by respondents who were either younger than 16 years old or

older than 25 years old. The data cleaning process performed by the partners followed the

common procedure outlined in this protocol.

Two rounds of data cleaning were performed by the western German team. During the first

round, the datasets were checked for interviewer errors. Code “99” was used for missing

values. The same procedures (checking typing errors, ensuring uniformity and correct

labelling) were applied for interviewer questionnaires and contact sheets. Each respondent

was assigned a unique ID number. Afterwards, the Bremen MYPLACE team checked datasets

for consistency, defined missing values and ensured correct labelling. Following MMU

requirements, different codes and labels were used for different types of missing values.

Overall, the data cleaning process was not as necessary for CAPI fieldwork, since the data

entry software was originally designed around the questionnaire and contained validation,

hence, data cleaning was, to a large extent, integrated in the data collection process.

3.7. Data analysis

MYPLACE August 2014


The first stage of data analysis consisted of national level analysis following the data analysis

template developed by MMU, in order to ensure similar analytic approaches. Frequency

distributions of all answers were checked, and a number of scales and indices were

constructed from multi-item response questions. For univariate analysis, the partners have

computed percentage, mean, median, and/or standard deviation for the variables under

analysis. For bivariate analysis, assumptions for parametric tests were checked. If

assumptions were fulfilled, Pearson Correlation, T-test and ANOVA were used. When the

assumptions were not fulfilled, non-parametric tests of Mann-Whitney, and Kruskal-Wallis

tests were used. For the analysis of two categorical variables, a Chi-square test was used

followed by Cramer's V/Phi (for nominal level variables) and Gamma (for ordinal level

variables).

The results of national level analyses are presented in Deliverable 4.5: A series of country

specific analysis which highlight local historical and cultural factors and which contrast the

two regions sampled.6

Cross-country analysis focused on the major themes studied by the MYPLACE project. It was

the second stage of WP4 data analysis, and is presented in the current Deliverable 4.6

report. Exploratory factor analysis was used to examine factor structure of the scale

measuring a construct. Cronbach Alpha was then used to test the reliability of scales or sub-

scales. Cronbach Alpha value of 0.60 was considered as the minimum value required for a

reliable scale in the analysis. A guidance document detailing this process can be found in

Appendix A: Guidance Document (preparing combined data for the thematic report)

Cross-country analysis of the following main substantial topics was performed at this stage:

Interest in politics;

Attitudes and trust towards social and political institutions;

Respondents’ political activism and civic engagement;

Opinions about effective political action;

Citizenship;

Social networks, gender and sexuality;

Religion;

Attitudes towards minority groups;

Democracy;

History and memory;

European issues.

6 http://www.fp7-myplace.eu/documents/WP4D4-5overviewreportv1.pdf.

http://www.fp7-myplace.eu/documents/WP4D4-5overviewreportv1.pdf

MYPLACE August 2014


A range of statistical techniques were used to analyse data. At a multivariate level, multi-

level modeling was mainly used, with the exception of Chapter 14 (History and Memory),

where multiple linear regression (OLS) was used.

3.8 Presentation of results

MYPLACE survey data were first presented in the WP4 deliverable report and MYPLACE

policy briefs. In these cases, statistical significance is on the basis of p values <0.05, although

in a number of cases it is higher, p < 0.01 or 0.005.

3.9 Sample characteristics

The tables presented in this section of the report provide information about demographic

characteristics of the respondents by survey locations.

Table 3.9.1. Respondents’ gender by survey location (%)

Location Male Female

Podsljeme (CRO) 53 47

Pescenica (CRO) 48 52

Odense East (DEN) 47 53

Odense Center (DEN) 48 52

Narva area (EST) 69 31

Tartu (EST) 49 51

Lieksa and Nurmes (FIN) 51 49

MYPLACE August 2014


Kuopio (FIN) 47 53

Kutaisi (GEO) 37 63

Telavi (GEO) 39 61

Bremen (GER-W) 51 49

Bremerhaven (GER-W) 54 46

Jena (GER-E) 46 54

Rostock (GER-E) 44 56

New Philadelphia (GRE) 58 42

Argyroupouli (GRE) 54 46

Sopron (HUN) 51 49

Ozd (HUN) 54 46

Agenskalns (LAV) 50 50

Forstate and Jaunbuve (LAV) 48 52

Lumiar (POR) 46 54

Barreiro (POR) 58 42

Kupchino (RUS) 47 53

Vyborg (RUS) 50 51

Rimavska Sobota (SLO) 48 52

Trnava (SLO) 49 51

Vic (SPA) 48 52

Sant Cugat (SPA) 49 51

Coventry (UK) 52 48

Nuneaton (UK) 45 55

All survey locations 50 50

Table 3.9.2. Respondents’ age by survey location (%)

Location 16 17 18 19 20 21 22 23 24 25

Podsljeme (CRO) 12 10 13 10 8 12 11 9 16 0

Pescenica (CRO) 10 7 10 9 9 11 13 11 19 1

Odense East (DEN) 3 7 9 8 8 8 17 16 12 11

Odense Center (DEN) 2 5 7 6 6 10 18 15 16 14

Narva area (EST) 5 7 5 3 5 10 13 14 20 18

Tartu (EST) 8 7 9 10 8 12 8 15 13 11

MYPLACE August 2014


Lieksa and Nurmes (FIN) 22 31 20 6 4 4 5 3 2 3

Kuopio (FIN) 11 8 6 8 9 10 13 12 12 11

Kutaisi (GEO) 8 11 8 10 12 12 14 11 11 3

Telavi (GEO) 12 11 10 10 10 12 10 7 10 7

Bremen (GER-W) 11 13 11 11 8 10 12 11 11 2

Bremerhaven (GER-W) 8 14 13 11 9 10 10 11 9 5

Jena (GER-E) 3 3 4 7 12 15 20 15 15 7

Rostock (GER-E) 5 5 5 6 9 11 16 18 15 8

New Philadelphia (GRE) 19 16 9 6 8 8 9 10 9 5

Argyroupouli (GRE) 19 12 7 8 10 8 12 11 11 4

Sopron (HUN) 4 10 11 10 13 12 11 11 12 8

Ozd (HUN) 5 9 10 11 11 13 13 14 9 6

Agenskalns (LAV) 9 12 10 10 10 8 13 10 11 8

Forstate and Jaunbuve (LAV) 17 14 9 10 10 8 7 9 10 8

Lumiar (POR) 10 11 11 9 8 12 12 11 7 9

Barreiro (POR) 15 15 14 13 9 7 6 5 6 10

Kupchino (RUS) 15 14 17 13 8 9 9 7 6 5

Vyborg (RUS) 16 9 11 8 12 10 10 8 10 7

Rimavska Sobota (SLO) 14 9 13 11 8 9 10 10 11 5

Trnava (SLO) 12 9 8 7 8 8 10 14 12 12

Vic (SPA) 8 9 10 9 11 9 10 10 12 13

Sant Cugat (SPA) 10 13 10 13 10 9 8 10 9 8

Coventry (UK) 8 9 6 16 16 13 11 8 8 6

Nuneaton (UK) 19 18 13 9 7 8 7 6 7 5

All survey locations 11 11 10 9 9 10 11 11 11 7

MYPLACE August 2014


Table 3.9.3. Respondents’ education by survey location (%)

Location

Less

th

an p

rim

ary

ed

uca

tio

n

Pri

mar

y e

du

cati

on

Low

er s

eco

nd

ary

ed

uca

tio

n

Up

per

sec

on

dar

y e

du

cati

on

Po

st-s

eco

nd

ary

no

n-t

ert

iary

edu

cati

on

Sho

rt-c

ycle

ter

tiar

y e

du

cati

on

Bac

hel

or’

s o

r eq

uiv

alen

t le

vel

Mas

ter’

s o

r e

qu

ival

ent

leve

l

Do

cto

ral o

r e

qu

ival

ent

leve

l

No

t el

sew

her

e cl

assi

fied

Podsljeme (CRO) 22 10 46 4 14 4

Pescenica (CRO) 1 18 12 46 4 14 4 1

Odense East (DEN) 1 29 54 4 11 1 1

Odense Center (DEN) 1 20 59 5 14 1 1

Narva area (EST) 2 24 11 20 32 11 1

Tartu (EST) 4 32 4 37 10 11 2

Lieksa and Nurmes (FIN) 2 75 18 4

Kuopio (FIN) 31 52 17

Kutaisi (GEO) 0 27 55 17

Telavi (GEO) 0 36 44 19 0

Bremen (GER-W) 0 27 33 11 0 28

Bremerhaven (GER-W) 1 33 27 12 0 27

Jena (GER-E) 10 63 15 1 10

Rostock (GER-E) 0 14 58 14 1 12

New Philadelphia (GRE) 2 39 47 2 1 9 1

Argyroupouli (GRE) 0 1 31 54 4 1 8 1 0

Sopron (HUN) 1 21 16 51 3 6 2 0 0

Ozd (HUN) 4 35 24 31 2 3 0 1

Agenskalns (LAV) 3 33 43 1 4 14 2

Forstate and Jaunbuve (LAV) 0 7 37 41 3 3 9 1

Lumiar (POR) 0 30 49 15 6 0

Barreiro (POR) 0 51 39 8 1 1

Kupchino (RUS) 1 19 51 12 14 2

Vyborg (RUS) 0 5 52 26 14 2

Rimavska Sobota (SLO) 1 42 14 38 1 3 1 1

Trnava (SLO) 25 13 44 1 9 4 4

Vic (SPA) 1 4 34 39 8 11 4

Sant Cugat (SPA) 0 4 27 48 5 13 2

Coventry (UK) 3 21 46 9 10 8 2

Nuneaton (UK) 8 43 23 19 5 0 1

All survey locations 1 9 25 44 3 2 11 2 0 3

MYPLACE August 2014


Table 3.9.4. Respondents’ citizenship by survey location (%)

Location Citizen Not a citizen






Tartu (EST) 99 1


Kuopio (FIN) 99 1

Kutaisi (GEO) 100 0

Telavi (GEO) 99 1

Bremen (GER-W) 91 9


Jena (GER-E) 93 7




Sopron (HUN) 99 1

Ozd (HUN) 100 0



Lumiar (POR) 97 3


Kupchino (RUS) 98 2

Vyborg (RUS) 99 1

Rimavska Sobota (SLO) 100

Trnava (SLO) 100 0

Vic (SPA) 80 20


Coventry (UK) 68 32

Nuneaton (UK) 97 3


MYPLACE August 2014


Table 3.9.5. Respondents’ country of birth by survey location (%)

Location

Born in the

country of

interview

Not born in the

country of

interview






Tartu (EST) 99 1


Kuopio (FIN) 98 2

Kutaisi (GEO) 99 1

Telavi (GEO) 98 2

Bremen (GER-W) 90 10


Jena (GER-E) 91 9




Sopron (HUN) 99 1

Ozd (HUN) 99 1



Lumiar (POR) 93 7


Kupchino (RUS) 93 7

Vyborg (RUS) 98 2


Trnava (SLO) 99 1

Vic (SPA) 77 23


Coventry (UK) 51 49

Nuneaton (UK) 93 7


MYPLACE August 2014


Table 3.9.6. Respondents’ nationality by survey location (%)

Location

National of

the country of

interview

Other

national






Tartu (EST) 97 3


Kuopio (FIN) 97 3

Kutaisi (GEO) 100 0

Telavi (GEO) 98 2

Bremen (GER-W) 91 9


Jena (GER-E) 93 7




Sopron (HUN) 99 1

Ozd (HUN) 100 0



Lumiar (POR) 93 7


Kupchino (RUS) 99 1

Vyborg (RUS) 99 1


Trnava (SLO) 100 0

Vic (SPA) 80 20


Coventry (UK) 59 41

Nuneaton (UK) 93 7


MYPLACE August 2014


Table 3.9.7. Respondents’ religion by survey location (%)

Location Catholic Protestant

Christian

Orthodox Muslim Jewish

Other

religion None

Podsljeme (CRO) 84 1 5 1 0 3 6

Pescenica (CRO) 79 1 1 5 2 11

Odense East (DEN) 2 52 19 0 10 17

Odense Center (DEN) 2 64 2 0 7 26

Narva area (EST) 1 5 73 1 1 19

Tartu (EST) 3 17 9 4 67

Lieksa and Nurmes (FIN) 85 5 1 1 8

Kuopio (FIN) 1 75 2 1 0 21

Kutaisi (GEO) 100 0

Telavi (GEO) 0 98 1 0 0

Bremen (GER-W) 15 50 2 10 0 4 20

Bremerhaven

(GER-W)

15 57 1 9

2 16

Jena (GER-E) 12 38 2 1 1 2 44

Rostock (GER-E) 9 30 1 1 0 4 55

New Philadelphia (GRE) 0 0 94 2 1 3

Argyroupouli (GRE) 0 0 94 1 2 3

Sopron (HUN) 70 16 1 13

Ozd (HUN) 52 11 8 4 24

Agenskalns (LAV) 20 25 19 0 0 3 34

Forstate and Jaunbuve

(LAV)

35 4 36

12 12

Lumiar (POR) 58 2 0 0 6 34

Barreiro (POR) 56 3 1 6 34

Kupchino (RUS) 1 1 70 3 1 8 17

Vyborg (RUS) 3 4 78 1 0 6 9

Rimavska Sobota (SLO) 55 19 1 25

Trnava (SLO) 73 7 0 0 19

Vic (SPA) 38 1 15 0 6 41

Sant Cugat (SPA) 45 1 1 0 2 50

Coventry (UK) 12 10 31 0 28 19

Nuneaton (UK) 13 9 13 23 43

All survey locations 26 17 25 4 0 5 23

MYPLACE August 2014


Table 3.9.8. Respondents’ Parental Social Class by survey location (%)

Location PC1: Low PC2 PC3 PC4: High

Podsljeme (CRO) 54 14 11 20

Pescenica (CRO) 58 16 10 16

Odense East (DEN) 51 10 16 22

Odense Center (DEN) 39 12 15 33

Narva area (EST) 67 12 10 11

Tartu (EST) 46 17 16 21

Lieksa and Nurmes (FIN) 64 18 10 8

Kuopio (FIN) 35 17 18 31

Kutaisi (GEO) 42 16 14 28

Telavi (GEO) 36 15 14 36

Bremen (GER-W) 51 14 13 23

Bremerhaven

(GER-W)

66 14 10 11

Jena (GER-E) 37 15 16 32

Rostock (GER-E) 41 16 14 30

New Philadelphia (GRE) 74 11 6 9

Argyroupouli (GRE) 63 10 11 15

Sopron (HUN) 70 5 11 14

Ozd (HUN) 93 2 4 2

Agenskalns (LAV) 51 10 14 24

Forstate and Jaunbuve

(LAV)

74 9 7 10

Lumiar (POR) 49 5 10 37

Barreiro (POR) 81 6 6 7

Kupchino (RUS) 29 13 18 39

Vyborg (RUS) 51 11 12 26

Rimavska Sobota (SLO) 83 6 7 4

Trnava (SLO) 67 11 11 11

Vic (SPA) 66 9 10 15

Sant Cugat (SPA) 30 11 16 42

Coventry (UK) 50 15 12 23

Nuneaton (UK) 70 14 11 5

All survey locations 56 12 12 20

MYPLACE August 2014


Chapter 4: Political Interest

This chapter examines the variables describing political interest. Respondents’ political

interest was measured using a range of questions, including: specific questions on

respondents’ interest in geographical, political and administrative units in different distance

from the respondent’s place of residence; and interest in ten themes related to public

affairs, social issues and topics and politics in general.

An overview of survey results and summary of national reports

In general, the level of political interest of young people is not very high: 42%, or slightly less

than half, consider themselves interested in politics and 58% report that they are not very

interested or not at all interested in politics.

Across countries, interest in politics differs. The Croatian report termed youth interest “very

low” since 70% reported ‘not at all’ or ‘not very interested’ in politics. Likewise the Latvian

report reported similar percentages as did the Finnish report. In contrast other countries –

eastern Germany, Greece, Spain, the political interest of youth was considered to be rather

high. Most of reports mentioned that interest depended on the concrete theme, which

respondents assessed. When young people were asked about their interest in five different

types of administrative/political/geographical units, the overall political interest was higher

and there was variation across the types of units.

MYPLACE August 2014


Graph 4a. How interested would you say you are in issues to do with ... % very interested

or quite interested.

Evidently, country and municipality are the main points of reference for the highest

percentage of respondents, whereas, Europe at large, and respondents’ immediate

community, were mentioned by a lesser percentage. High interest in one’s country was

mentioned in a several reports (eastern and western Germany, Slovakia, Estonia). In the

report from Russia, we find that youth were more interested in immediate surroundings.

Regarding interest in social issues pertaining to and influencing young people’s lives directly,

one can see that some issues clearly differentiated interest (the issue most interesting to

respondents has a mean value two times larger than the issue least interesting for them)

while other issues were of similar interest.

MYPLACE August 2014


Graph 4b Please tell me to what extent you are interested in each of the following issues,

on a scale of 0 to 10? To what extent are you personally interested in… Mean values on

scale from 0 (not at all) … 10 (very interested).

General employment prospects and housing were the two issues that aroused most

interest. In addition, there was quite a lot of interest in environmental issues. LGBT rights

were interesting for the least percentage of youth. Interestingly immigration and EU issues

scored low. Obviously high interest in employment and housing is a consequence of the

relevance of those spheres for youth in general (though not for every young person in the

age group, 16-25 years). High interest in these issues was mentioned in several reports

(including, Croatia, Estonia, Portugal, and Russia).

Regarding information channels from where young people obtained information about

politics and public affair, television and the Internet were notably more popular than

newspapers and radio. The difference is even more marked when we compare percentages

of those who reported following politics at least 30 minutes a day.

MYPLACE August 2014


Graph 4c On an average day, how much time do you spend keeping yourself informed

about politics and current affairs, using the following media?

Croatia, eastern Germany, Estonia, Portugal and Spain reported roughly similar patterns of

mass media channel usage for obtaining information about politics and public affairs.

Literature

In an extensive study of the origins of political interest, Shani (2009) defines her object of

research as‘… political interest, defined as the intrinsic motivation to engage in politics…’.

Political interest is understood as the driving force behind the search for information,

behind choices on what to follow on news and TV broadcasts and what makes a person

focus on public matters and politics. It is the factor, which significantly influences how

people get involved in different forms of political participation. Political interest is perhaps

the strongest single influence on political knowledge and the ability to reason about politics.

In addition, it influences also a myriad of political attitudes, beliefs and understandings

(Shani 2009: 2-24).

Van Deth, (1990) tackles the question of the duality of political interest. Is political interest

something pertaining to the inner mental world only or does it have an external behavioural

component too? For the sake of analytical clarity his choice is to define interest as ‘the

degree to which politics arouses a citzen’s curiosity‘(1990: 278). This wording emphasises

the mental nature of political interest, without seeing it having behavioural components.

MYPLACE August 2014


Holleque (2011) makes the distinction between interest as a mental state and actual

behaviour. For him, political interest is attention to politics. Attention to politics means

readiness to inform oneself, as well as readiness to participate. However, it is not a

behaviour itself, whether information seeking behaviour or political action. The argument

against seeing political interest as comprised of attitudes and behaviour maintains that all

types of behaviours are influenced by a range of factors and political interest, at best, is only

one factor among others, but it might be not present at all. In addition, Niemi, Craig and

Mattei (1991: 1410) have made a distinction between three aspects of psychological

involvement in politics when they talk about general interest in politics, namely discussing

politics and following politics on TV. Therefore, they too opt to define political interest as a

feature of a mind-set only, without behavioural aspects being present. Elsewhere too

political interest appears in the section devoted to studying political attitudes. It has been

measured using only one question, ‘How interested would you say you personally are in

politics?’ (Utter, 2011). Thus political interest is understood as a general predisposition of an

individual, not as a concrete form of behaviour. Syal (2012: 423) has made a distinction

between ‘general interest of citizens in public affairs, apart from times of actual elections‘

which is continuous and ‘political participation’, which refers to certain acts and activities in

connection with elections. Here too we see that behaviour and interest are considered two

different things.

Lange et al. (2013) differentiate between long-term and situational term interest in politics.

While the first refers to ‘relatively stable preferences to certain issues, which is paid certain

attention in the long-term‘, the other one is linked to a certain situation. It is possible that

long-term interest gets started from a situational interest. For them, interest represents also

the value orientations of individuals or the readiness to act in a certain manner. Thus it is a

mental quality, not a behaviour or action, even if it constitutes a precondition for

information seeking or a concrete act. Schmidt (2004: 14) defined political interest as

‘curiosity, relation, attention and vigilance to political issues‘ which conforms with the

understandings that political interest is rather a state of mind than action. Also Niedermayer

(2005) and Hadjar and Becker (2006) define interest in politics as a precondition for political

activities but not activities themselves.

Kuhn and Schmidt (2004: 75) distinguish between different spheres of interest:

1. Classic policies (e.g. policy of international organisations– NATO, UNO, EU, foreign policy, domestic policy, policy of parties, defence policy, economic policy),

2. Social policies (e.g. labour market policy, social policy, health policy, retirement pension policy),

3. New policies (e.g. policy towards third world countries, peace policy, environment protection policy)

4. Migration policies (e.g. immigration policy, integration policy). Kenski and Stroud (2012: 180-181) locate political interest among attitudes and political

identification, but operationalise it as activities, albeit activities which do not require very

big effort: including, how often people follow what is going on in government and public

MYPLACE August 2014


affairs, or how many days in the past week they had talked with family and friends about

politics. This was however differentiated from following specific political campaigns, such as

presidential elections, on traditional media channels (TV, newspapers, radio). Thus, here we

see that political interest is understood as a set of behaviours. Pritzker et al. (2012: 6-10)

have a different view on political interest. Their empirical approach is embedded in the

political participation approach and thus the index of political interest is constituted mainly

of activities, which in other studies, including MYPLACE, appear as forms of participation.

The only non-activism variable which is in the questionnaire is ‘volunteering in an election

campaign’.

Factors, which influence the level of political interest

In the political activism and political participation literature, political interest appears mostly

as an independent variable, as a factor, which influences other variables. The list of variables

that are used range from political knowledge, stability of attitudes to political action.

It has been found that the level of self-assessed and self-reported political interest is highly

stable over time, both in short- (a year) and long-term (decades) periods (Prior, 2010). This

finding has two important implications. First, an individual’s political interest remains

roughly the same throughout his or her life course (at least over several decades). However,

the age from which levels of political interest stay stable, probably is not clear. Second, only

the most unique events have the capability of influencing a person’s level of interest in

politics. For example, the Fall of the Berlin Wall / reunification of Germany, could only

increase interest in politics in the short term perspective.

In the long-term perspective, political interest seems to be a stable mental construct, ‘a kind

of unmovable mover‘. Its (initial) levels are set by parents and members of one’s family

during childhood and youth, through discussing public affairs, politics, community matters,

as well as through other experiences at home (see Shani, 2009; Verba, Schlozman and

Brady, 1995: 459; Tenscher and Scherer, 2012). Its levels can, vary significantly, depending

on the situation and its relevance to an individual. Among socio-economic variables, three

variables have been found to significantly influence the level of political interest: the level of

education; gender; and age. However, the explanatory power of these three variables is

rather weak (van Deth, 1990: 301-307).

Another variable that is strongly linked to the level of interest is ‘activism at school and in

NGOs’ (Verba et al.,1995: 434). Unfortunately, it is impossible to ascertain the causal

direction of influence or even be sure of what would be the accurate causal model. We

cannot be sure whether participation increases interest or is it vice versa; or might family

background, for example, influence both of them. Tenscher and Scherer (2012) report a

relationship between political knowledge and political interest, but it is impossible to

ascertain causality between these two, or which way does causal influence run. Schmid

(2004: 18,19) found that people characterised by ‘higher interest’ are capable of placing

MYPLACE August 2014


themselves on a left-right scale and on scales of political attitudes – in other words they are

characterised by a better understanding of politics, public affairs and politics. Nevertheless,

we do not know which comes first, political interest or capability to orient towards social

and political issues.

One interesting feature of political interest is the influence of political interest on

conventional political activities, especially voting and related activities, but not so in the

case of protest activities and other similar activities. While studies of conventional political

activism routinely report positive relationships between levels of political interest and

participation (Verba et al.,1995: 334-369), one large scale study of political protest hardly

even mentions political interest and does not consider it among key variables explaining

involvement in protest activities (Kaase and Marsh, 1979: 41-50; Kaase, 1990).

Summary and key issues for the MYPLACE analysis

This short review of political interest carries three clear issues relevant for the analysis of

political interest in the MYPLACE project. First, it is necessary to make a clear distinction

between mental predispositions – curiosity, interest, motivation or something else – and a

range of behaviours, which might be motivated, and initiated by this mental predisposition.

Our choice at this point is to concentrate on that mental predisposition only, and leave

information-seeking behaviour out. The rationale behind doing so is that the spectrum of

concepts and variables, which vary together with the level of political interest, is very wide.

It certainly is not limited to talking to others and/or following news about politics. Thus,

there is no particular reason to view political discussion and following mass media as

integral aspects of political interest. It would be more appropriate to view the latter as

aspects of political activism and information acquisition behaviours, not as political interest

even though they have close link to political interest. Second, political interest is not

internally structured around certain themes. There is no good reason to distinguish between

specific political interest in foreign relations, in employment or in minorities’ rights. Though

these, and also other themes probably are not addressed by the same individuals, it seems

that the motivation behind becoming and being interested in particular themes is largely the

same for all individuals. Third, in the search for the factors which influence the level of

political interest, the standard set of socio-demographic variables most probably is

sufficient. As well as socio-demograpchic predictors of political atttitudes there are also

other attitudinal and behavioural factors, which co-vary with the level of political interest,

but it is hard, if not impossible, to build an appropriate causal model, or to ascertain the

direction of causal flow, in such models. For instance, we cannot be sure whether political

interest influences organisational activism, or vice versa, or does a third factor (like parental

influence and family background) influence both of them.

MYPLACE August 2014


Items used to measure political interest

Based on the reasons explained in the theoretical section and due to the needs of data

analysis the questionnaire items on political interest were divided into two groups:

a) items measuring behavioural aspects and respondents’ assessments of their close relatives’ and friends’ interest in politics measuring

b) self-assessed level of political interest – items measuring non-behavioural that is purely mental aspects of political interest.

The selection of variables for data analysis is influenced also by additional conditions.

Question2 and 4 items ask about ‘political interest of and interaction with respondent’s

close friends and relatives’. Since not every respondent has each of the referred persons

present in the household, there is a lot of variation, which results from family composition

and not from political interest of the respondent. We do not account for variation, which is

caused by the household composition or friendship circles. Questions which contain

variation, which is not present in questions focusing on the self only, need be analysed

separately from the questions dealing with the self only.

For measuring mental aspects of political interest, questions from three items sets in the

questionnaire were chosen: Q1, Q2, Q5

Based on the results from the Factor and reliability analysis, it was decided to develop two

cumulative indices and to analyse three variables individually:

Index1: Interest in different geographical regions (Q1_1 … Q1_5),

Index2: Interest in a range of topics (Q5_3 … Q5_10),

4.1 Interest in areas (Q1)

Interest in areas include the following question and variables;

Q1 How interested would you say you are in issues to do with …

Q1_1 ... the neighbourhood you currently live in?

Q1_2... the city you currently live in?

Q1_3 ... the country you live in?

Q1_4 ... countries neighbouring /country/?

Q1_5 ... Europe?

Respondents could choose from ‘very interested’, ‘quite interested’, ‘not very interested’

and ‘not at all interested’

Exploratory factor analysis revealed the presence of an internal structure in the responses

when items in Q1 are examined. The reliability of the scale, consisting of items involved in

Factor 1, was high, the Cronbach alpha was 0.807. Across countries, the value varied from

0.596 in Denmark to 0.890 in Hungary.

MYPLACE August 2014


Q1: Interest in areas (cumulative index)

Variable Description

Dimensionality and reliability analysis presented above suggested that

the five items make a good one-dimensional cumulative scale.

The cumulative index has a minimum value 0, indicating no interest in

any of the five units and a maximum value 15, indicating maximum

possible interest in the five units. The index sums up responses from

16,733 valid cases and has 202 missing cases (1.2%).

Graph 4.1: Mean interest in issues to do with different geographical and

administrative units (cumulative index) by location

Narrative

The overall mean of the index is 8.81 and the standard deviation is 3.04.

This varies by location, with the highest mean score at Kakhinauri region

of Kutaisi with 10.58, and the lowest mean score at Jauna Forstadte in

Daugavpils (Latvia) with 6.73.

There are no clear clusters of locations. Nevertheless, three locations

with the highest level of interest can be pointed out: Telavi and

Kakhinauri (Georgia) and Telheiras (Portugal). Six locations with the

lowest levels of interest are: Trnava and Rimavska Sobota (Slovakia),

Vyborg (Russia), Nuneaton (UK), Pescenica Zitnjak (Croatia) and Jauna

Forstande (Latvia).

Most Countries had a similar mean score between the two locations.

Still, two groups could be distinguished which had higher differences: in

Portugal, western Germany and Finland the differences were

respectively 0.73, 0.77 and 0.91. In Russia, the UK and Latvia the

differences were 1.49, 1.64 and 1.98.

MYPLACE August 2014


4.2 Interest in politics (Q2)

Question 2 asked the young person ‘How interested do you think the following people are in

politics?’ for the following key individuals;

Q1_1 You, Yourself

Q2_2 Your father

Q2_3 Your mother

Q2_4 The brother or sister that you are closest to

Q2_5 The grandparent that you are closest to

Q2_6 Your boyfriend / girlfriend / partner

Q2_7 Your best friend

The following response options were; ‘very interested’, ‘quite interested’ , ‘not very

interested’ and ‘not at all interested’.

Responses to these questions are significantly influenced by the composition of the

respondent’s household. The table below presents percentage of missing values, per

question, across countries where one can see that in some combinations, more than half of

the sample has not checked an answer to a question. Building a cumulative index would

either require replacement of missing values or result in a composite variable, which has

missing values for most cases. Option one would require a general strategy which would

outline how to replace missing values, which has not been adopted for MYPLACE deliverable

D4.6. Option two is not plausible since it would probably lead to biased estimates because

the absence of certain family members is probably not completely random, but to some

significant extent systematic and related to the household’s socio-economic background.

Since either option is not a plausible solution, we decided to analyse the variables

individually.

MYPLACE August 2014


Table 4.2 Percentage of missing values per country and question. The question: How

interested do you think the following people are in politics?

Country Your father

Your

mother

The brother

or sister

that you are

closest to

The

grandparen

t that you

are closest

to

Your

boyfriend /

girlfriend /

partner

Your best

friend

Croatia 9 7 20 26 42 9

Denmark 5 2 5 20 44 2

Estonia 22 6 32 32 35 8

Finland 12 8 18 27 48 9

Georgia 17 5 22 31 51 4

Western

Germany 10 4 21 24 48 2

Eastern Germany 8 1 18 13 36 2

Greece 5 3 27 32 42 4

Hungary 14 7 17 26 37 12

Latvia 17 5 36 25 30 3

Portugal 11 6 35 30 39 10

Russia 21 9 27 20 24 9

Slovakia 9 4 20 20 33 12

Spain 9 4 28 26 52 2

UK 12 4 14 20 44 3

Average 12% 5% 23% 25% 40% 6%

MYPLACE August 2014


Q2_1: Personal interest in politics?


The variable has a minimum value 1, indicating high interest in

politics and a maximum value 4, indicating low interest in politics.

The variable has 16,843 valid cases and has 92 missing cases (0.5%).

Graph 4.2a: Self-assessed interest in politics. Percentages of response

categories by location

Narrative

The overall mean of the variable is 2.66 and the standard deviation is

0.89. This varies by location, with the lowest mean value / highest

interest at Jena (eastern Germany) with 2.15, and the highest mean

value / lowest interest at Sopron (Hungary) with 3.10.

There are no clear clusters of locations. Three locations with the

highest mean value / lowest interest in politics are Sopron and Ozd

(Hungary) and Forstate & Jaunbuve (Latvia). Two locations with the

lowest mean value / highest interest in politics are Jena and Rostock

in eastern Germany.

MYPLACE August 2014


Q2_2: How interested do you think the following people are in politics? Your father


The variable has a minimum value 1, indicating high interest in politics

and maximum value 4, indicating low interest in politics.

The variable has 14,865 valid cases and has 2,070 missing cases

(12.1%). Most of the missing cases are ‘does not apply’ responses

(7.8%), ‘don’t know’ was checked by 3.8% or respondents.

Graph 4.2b: Respondent-assessed father’s interest in politics.

Percentages of response categories by location

Narrative


0.88. This varies somewhat by location, but not greatly, with the

lowest mean value / highest interest at Odense Centre (Denmark)

with 1.83, and the highest mean value / lowest interest Sopron

(Hungary) with 2.60.

There are no clear clusters of locations. Three locations with the

highest mean value / lowest interest in politics are Sopron and Ozd in

Hungary and Narva in Estonia. The three locations with the lowest

mean value / highest interest in politics are Odense Centre and

Odense East in Denmark and Sant Cugat in Spain.

The UK had the largest differences between locations (0.41), Greece

had the second largest (0.28), and Portugal third largest (0.27).

Georgia and Denmark had the smallest difference (0.01).

MYPLACE August 2014


Q2_3: How interested do you think the following people are in politics? Your mother


The variable has a minimum value 1, indicating high interest in politics and a

maximum value 4, indicating low interest in politics.

The variable has 16,098 valid cases and has 837 missing cases (4.9%). Most

of the missing cases are ‘don’t know’ (2.5%), ‘does not apply’ was checked

by 2.0% of respondents.

Graph 4.2c: Respondent-assessed mother’s interest in politics.


Narrative

The overall mean of the variable is 2.47 and the standard deviation is 0.86.

This varies somewhat by location, but not by much, with the lowest mean

value / highest interest at Odense East (Denmark) with 2.13, and the highest

mean value / lowest interest Jauna Forstadte (Latvia) with 2.80.

There are no clear clusters of locations. The three locations with the highest

mean value / lowest interest in politics are Jauna Forstadte in Latvia, and

Sopron and Ozd in Hungary. The three locations with the lowest mean value

/ highest interest in politics are Odense Centre and Odense East in Denmark

and Sant Cugat in Spain.

Portugal had the largest differences between locations (0.38), Spain had the

second largest (0.30), Latvia third largest (0.24) and Russia fourth largest

(0.20). Other countries had a difference smaller than 0.15.

MYPLACE August 2014


Q2_4: How interested do you think the following people are in politics? Your brother or sister you are closest to



and a maximum value 4, indicating low interest in politics.


(23.3%). Most of the missing cases are ‘does not apply’ (18.6%), ‘don’t

know’ was checked by 3.5% of respondents.

Graph 4.2d: Respondent-assessed brother’s or sister’s interest in

politics. Percentages of response categories by location

Narrative



interest at Odense East (Denmark) with 2.53, and the highest mean

value / lowest interest Jauna Forstadte (Latvia) with 3.32.

There are no clear clusters of locations. The three locations with the

highest mean value / lowest interest in politics are Jauna Forstadte in

Latvia, and Sopron and Ozd in Hungary. The three locations with the

lowest mean value / highest interest in politics are Odense Centre and

Odense East in Denmark and Bremen in western Germany.

Portugal had the largest differences between locations (0.48). Latvia

had the second largest difference (0.26) and the UK third largest

(0.25). eastern Germany, Spain, Croatia, Denmark and Hungary had

differences smaller than 0.10.

MYPLACE August 2014


Q2_5: How interested do you think the following people are in politics? The grandparent that you are closest to






know’ was checked by 6% of respondents.

Graph 4.2e: Respondent-assessed grandparent’s interest in politics.


Narrative



interest at Telavi (Georgia) with 2.04, and the highest mean value /

lowest interest Ozd (Hungary) with 2.93.

The two locations with the highest mean value / lowest interest in

politics are Ozd in Hungary and Barreiro in Portugal. The location with

the lowest mean value / highest interest in politics is Telavi in Georgia.

Portugal had the largest differences between locations (0.35). Spain

and Estonia had the second largest difference (0.28). Finland,

Hungary, Russia, Georgia had the difference between 0.20 and 0.25.

MYPLACE August 2014


Q2_6: How interested do you think the following people are in politics? Your boyfriend / girlfriend / partner






know’ was checked by 2.8% of respondents.

Graph 4.2f: Respondent-assessed partner’s interest in politics.


Narrative



interest at Odense centre (Denmark) with 2.18, and the highest mean

value / lowest interest Jauna Forstadte (Latvia) with 3.12.

The two locations with the highest mean value / lowest interest in

politics are Jauna Forstadte in Latvia and Sopron in Hungary. The

location with the lowest mean value / highest interest in politics is

Odense Centre in Denmark.

Portugal had the largest differences between locations (0.48). Georgia

had the difference 0.37, Finland 0.34. Other countries had the

difference around 0.2 or below that.

MYPLACE August 2014


Q2_7: How interested do you think the following people are in politics? Your best friend


The variable has a minimum value 1, indicating high interest in

politics and a maximum value 4, indicating low interest in politics.

The variable has 15894 valid cases and has 1041 missing cases

(6.1%). Most of the missing cases are ‘don’t know’ (3.9%), ‘does not

apply’ was checked by 1.6% of respondents.

Graph 4.2g: Respondent-assessed best-friend’s interest in politics.


Narrative

The overall mean of the variable is 2.72 and the standard deviation

is 0.90. This varies by location, with the lowest mean value / highest

interest at Odense East (Denmark) with 2.22, and the highest mean

value / lowest interest Ozd (Hungary) with 3.17.

Sopron and Ozd in Hungary and Jauna Forstadte in Latvia are the

three locations with the highest mean value / lowest interest in

politics. The two locations with the lowest mean value / highest

interest in politics are both locations of Odense in Denmark.

Portugal had the largest differences between locations (0.48). The

UK had the difference 0.40 and Finland 0.23. Other countries had

the difference around 0.16 or below that.

MYPLACE August 2014


4.3 Time use in media for politics and current affairs (Q3)

Q3 On an average day, how much time do you spend keeping yourself informed about politics and

current affairs using the following media?

Q3_1 The radio

Q3_2 The internet

Q3_3 TV

Q3_4 Newspapers

Response options; ‘no time at all’, ‘less than 1/2 hour’, ‘1/2 hour to 1 hour’ and ‘ more than

1 hour’.

Dimensionality analysis, using exploratory Factor analysis, returned a single factor in the

pooled dataset, as well as in all individual countries. This is an indication that responses to

the four questions are to a notable degree, positively correlated. To assess if this correlation

is sufficient, we conducted reliability analysis. The reliability analysis returned values of

Cronbach alpha that were too low to develop a cumulative index. Even if there are six

countries where the value of the statistic is sufficiently high, the overall value (0.630) is too

low; and it is too low in most of the countries. The values of the individual Cronbach alpha

ranged from 0.783 in Slovakia to 0.343 in eastern Germany. Since there are no good

empirical grounds for developing a cumulative index, we will analyse the items individually.

MYPLACE August 2014


Q3_1: Time spent keeping yourself informed about politics and current affairs using the following media: The Radio



The distribution does not come from a normal distribution.

Graph 4.3a: Time spent on keeping oneself informed about politics

using radio. Percentages of response categories by location

Narrative


0.88. This varies by location, with the lowest mean value / less time at

Barreiro (Portugal) with 1.30, and the highest mean value / most time

at Rostock (eastern Germany) with 2.14.

Rostock and Bremerharven respectively in eastern and western

Germany are the two locations with the highest mean value / most

time spent of following politics. The three locations with the lowest

mean value / least time are both locations in Georgia and Barreiro in

Portugal.

Portugal had the largest differences between locations (0.37). Latvia

had the difference 0.29, Hungary 0.25 and eastern Germany 0.23. The

differences were the smallest in Estonia (0.00), Slovakia (0.03) and

Georgia (0.03).

MYPLACE August 2014


Q3_2: Time spent keeping yourself informed about politics and current affairs using the following media: The Internet



The distribution does not come from a normal distribution.

Graph 4.3b: Time spent on keeping oneself informed about politics

using internet. Percentages of response categories by location

Narrative


1.03. This varies by location, with the lowest mean value / lowest time

spent on getting political new in internet at Ozd (Hungary) with 1.67,

and the highest mean value Jena (eastern Germany) with 2.87.

Jena in eastern Germany is the location with the highest mean value.

The two locations with the lowest mean value Ozd in Hungary and

Barreiro in Portugal.

There is a clear cluster of countries, where the differences are large:

Portugal (0.52), Latvia (0.50), the UK (0.48) and Russia (0.42). Three

countries had nearly no difference: western Germany and Denmark

(0.01), Spain (0.02).

MYPLACE August 2014


Q3_3: Time spent keeping yourself informed about politics and current affairs using the following media: TV


The variable has 16,829 valid cases and has 106 missing cases (0.9%)

Graph 4.3c: Time spent on keeping oneself informed about politics

using TV. Percentages of response categories by location

Narrative


0.95. This varies by location, with the lowest mean value / least time

spent on following political news in TV at Jena (eastern Germany)

with 1.85, and the highest mean value at Telavi in Georgia with 2.85.

Both locations in Georgia are the two locations with the highest mean

value. Jena in eastern Germany is the location with the lowest mean

value.

MYPLACE August 2014


Q3_4: Time spent keeping yourself informed about politics and current affairs using the following media: Newspapers



Graph 4.3d: Time spent on keeping oneself informed about politics

using newspapers. Percentages of response categories by location

Narrative


0.76. This varies by location, with the lowest mean value / least time

on reading newspapers at New Philadelphia in Greece with 1.23, and

the highest mean value / most time at Lieksa and Nurmes (Finland)

with 1.89.

Two locations in Finland are the locations with the highest mean value

/ most time spent on reading newspapers. The three locations with

the lowest mean value / least time spent on newspapers are both

locations in Greece and in Georgia.

Portugal had the largest differences between locations (0.28) and

Hungary had the difference 0.26. Other countries had the difference

below 0.20. The three smallest differences were in Georgia, Russia

and Spain respectively 0.01, 0.02 and 0.03.

MYPLACE August 2014


4.4 Discussion on political issues with key individuals (Q4)

Question 4 asked the respondent ‘How often do you discuss political issues when you get

together with the following people?’ for a range of key individuals;

Q4_1 Your father

Q4_2 Your mother

Q4_3 The brother or sister that you are closest to

Q4_4 The grandparent that you are closest to

Q4_5 Your boyfriend / girlfriend / partner

Q4_6 Your best friend

There were five response options; ‘always’, ‘often’, ‘sometimes’, ‘rarely’ and ‘never.’

Responses to these questions are significantly influenced by the composition of

respondent’s household. The table below presents percentage of missing values, per

question, across countries where one can see that in some combinations, approximately

half of the sample has not checked an answer to a question. Building a cumulative index

would either require replacement of missing values or result in a composite variable, which

has missing values for most cases. Option one would require a general strategy which would

outline how to replace missing values, which has not been adopted for MYPLACE deliverable

D4.6. Option two is not plausible since it would probably lead to biased estimates because

the absence of certain family members is not completely random but to some significant

extent systematic and related to household’s socio-economic background. Since either

option is not a plausible solution, we decided to analyse the variables individually.

MYPLACE August 2014


Table 4.4 Percentage of missing values per country and question. The question: How often

do you discuss political issues when you get together with the following people?

Country Your father Your mother

The brother

or sister that

you are

closest to

The grandparent

that you are

closest to

Your

boyfriend /

girlfriend /

partner

Your best

friend

Croatia 6 3 16 19 38 3

Denmark 4 1 5 18 43 1

Estonia 19 4 28 25 34 5

Finland 5 2 8 14 43 2

Georgia 13 2 19 27 49 0

Western

Germany 6 2 19 20 48 1

Eastern Germany 6 1 17 11 35 1

Greece 3 1 26 27 39 1

Hungary 9 4 13 18 31 6

Latvia 15 3 35 24 30 2

Portugal 8 3 33 25 37 6

Russia 18 7 23 17 21 4

Slovakia 7 2 20 17 32 8

Spain 7 2 25 24 51 1

UK 9 3 12 17 43 1

Average 9% 3% 20% 20% 38% 3%

MYPLACE August 2014


Q4_1: How often do you discuss political issues when you get together with the following people? Your father



(9.4%). The ‘Does not apply’ response option was checked by 7.9%

respondents, and ‘don’t know’ by 0.5% respondents.

Graph 4.4a: Frequency of discussing politics with father. Percentages

of response categories by location

Narrative



frequency of talking with father about politics being reported at Sant

Cugat in Spain with 2.99, and the highest mean value / lowest

frequency at Sopron in Hungary with 4.03.

Narva in Estonia had the second highest mean value / lowest

frequency of talking about politics with father. Two locations in

addition to Sant Cugat where mean value was low / reported

frequency was high, were Odense Centre in Denmark and Vic in Spain.

The UK had the largest difference between sites (0.50) and Portugal

had the second largest differences between locations (0.39). Finland,

Slovakia, Estonia and Spain had the difference between 0.20 and 0.30.

Russia, Georgia, Greece, eastern Germany, Croatia, Denmark, Latvia

had the difference smaller than 0.10.

MYPLACE August 2014


Q4_2: How often do you discuss political issues when you get together with the following people? Your mother


The variable has 16468 valid cases and has 467 missing cases (2.8%).

The ‘Does not apply’ response option was checked by 2.0%

respondents, and ‘don’t know’ by 0.3% respondents

Graph 4.4b: Frequency of discussiing politics with mother.


Narrative



frequency of talking with mother about politics being reported at Sant

Cugat in Spain with 3.25, and the highest mean value / lowest


Jauna Forstadte in Latvia had the second highest mean value / lowest

frequency of talking about politics with mother. Odense Centre in

Denmark had the second lowest mean value / second highest

reported frequency.

Portugal had the largest difference between sites (0.46), and the UK

had second largest differences between locations (0.34). Georgia had

the difference 0.28, Spain 0.26, Finland 0.24 and Slovakia 0.23.

Estonia, Denmark, Greece and Croatia had the difference smaller than

0.08.

MYPLACE August 2014


Q4_3: How often do you discuss political issues when you get together with the following people?: The brother or sister that you are closest

to





Graph 4.4c: Frequency of discussing politics with brother/sister.


Narrative



frequency of talking with brother/sister about politics being reported

at Vyborg in Russia with 3.78 and the highest mean value / lowest

frequency at Jauna Forstadte in Latvia with 4.43.

Sopron in Hungary had the second highest mean value / lowest

frequency of talking about politics with brother/sister. Odense East in

Denmark had the second lowest mean value / second highest

reported frequency.

The UK had the largest difference between sites (0.49), and Portugal

had second largest differences between locations (0.41). Slovakia had

the difference 0.31 and Russia 0.27. Finland, Georgia, western

Germany, Latvia and Estonia had the difference between 0.20 and

0.10, and other countries below 0.10.

MYPLACE August 2014


Q4_4: How often do you discuss political issues when you get together with the following people? The grandparent that you are closest to





Graph 4.4d: Frequency of discussing politics with grandparents.


Narrative



frequency of talking with a grandparent about politics being reported

at Vyborg in Russia with 3.67 and the highest mean value / lowest

frequency at Ozd in Hungary with 4.37.

Lieksa and Nurmes in Finland, Barreiro in Portugal and Sopron in

Hungary had also high mean value / low frequency of talking about

politics with a grandparent. Rostock in eastern Germany and

Kuptchino in Russia had also lowest mean value / high reported

frequency.

Portugal had the largest difference between sites (0.35). Georgia had

the difference 0.23, and Finland had the difference 0.22. Eastern

Germany had the difference 0.13 and Spain 0.14. Other countries had

the difference 0.10 or smaller.

MYPLACE August 2014


Q4_5: How often do you discuss political issues when you get together with the following people? our boyfriend / girlfriend / partner





Graph 4.4e: Frequency of discussing poltics with partner.


Narrative



frequency of talking with partner about politics being reported at

Rostock in eastern Germany with 2.95 and the highest mean value /

lowest frequency at Sopron in Hungary with 4.31.

Jauna Forstadte in Latvia had the second highest mean value / lowest

frequency of talking about politics with partner. Jena in eastern

Germany had the second lowest mean value / second highest

reported frequency.

Finland had by far the largest difference between locations (0.70).

Georgia had the second largest difference between locations (0.46),

and Portugal the third largest difference (0.41). Spain, Denmark,

eastern Germany, Croatia, Russia and Greece had a difference smaller

than 0.10.

MYPLACE August 2014


Q4_6: How often do you discuss political issues when you get together with the following people? Your best friend



The ‘Does not apply’ response option was checked by 1.4%


Graph 4.4f: Frequency of discussing politics with best friend.


Narrative



frequency of talking with best friend about politics being reported at

Sant Cugat in Spain with 3.00 and the highest mean value / lowest


Ozd in Hungary had the second highest mean value / lowest

frequency of talking about politics with best friend. Jena in eastern

Germany had the second lowest mean value / second highest

reported frequency.

The UK had the largest difference between locations (0.56), Portugal

had the second largest differences between locations (0.46), Finland

had the third difference 0.40 and Georgia had the difference 0.36.

Latvia, Croatia, eastern Germany, Denmark, Hungary and Greece had

the difference 0.10 or below that.

MYPLACE August 2014


4.5 Interest in contemporary issues (Q5)

The following series of questions were asked using a 11 point scale (0 not interested … 10

very interested) To what extent are you personally interested in...?

Q5_1 immigration

Q5_2 LGBT rights (lesbian, gay, bisexual, and transgender)

Q5_3 the economy

Q5_4 national security

Q5_5 the environment

Q5_6 the European Union

Q5_7 general employment prospects

Q5_8 housing

Q5_9 local issues

Q5_10 women’s rights

A factor analysis identified that questions Q5_3 to Q5_10 could be used to form a good one-

dimensional cumulative scale. The Cronbach alpha was 0.887. Across countries, the value

varied from 0.713 in eastern Germany to 0.922 in Finland.

Therefore questions on Q5_1 immigration, and Q5_2 LGBT rights (lesbian, gay, bisexual,

and transgender) are analyses separately.

MYPLACE August 2014


Q5: Interest in a range of contemporary issues (Q5_3, Q5_4, Q5_5, Q5_6, Q5_7, Q5_8, Q5_9 and Q5_10)


Dimensionality and reliability analysis presented above suggested

that the eight items make a good one-dimensional cumulative scale.

The cumulative index has a minimum value 0, indicating no interest

in any of the eight issues and a maximum value 80, indicating

maximum possible interest in the eight themes.

The index sums up responses from 16,504 valid cases and has 431

missing cases (2.5%).

Graph 4.5a: Mean interest in a range of issues (cumulative index) by

location

Narrative

The overall mean of the index is 50.30 and the standard deviation is

16.56 This varies by location, with the highest mean score at Telavi

(Georgia) with 68.21, and the lowest mean score at Vyborg (Russia)

with 36.06.

There are no clear clusters of locations. Nevertheless, two locations

with the highest level of interest can be pointed out: Telavi and

Kutaisi (Georgia). Two locations with the lowest levels of interest are

Trnava (Slovakia) and Vyborg (Russia).

Most countries had a similar mean score between the two locations.

Still, Hungary had a notably higher difference, of 15.51. Differences

were larger also in Estonia (8.55), Latvia (7.78), Finland (7.61),

Slovakia (6.35) and the UK (6.00).

MYPLACE August 2014


Q5_1: Interest in Immigration.


The variable has a minimum value 0, indicating no interest in immigration

and a maximum value 10, indicating maximum possible interest in

immigration.


Graph 4.5b: Interest in immigrantion issues by location (mean values)

Narrative

The overall mean of the index is 5.27 and the standard deviation is 2.86.

This varies by location, with the highest mean score at New Philadelphia

(Greece) with 7.05, and the lowest mean score at Rimavska Sobota

(Slovakia) with 2.94. The differences across sites are quite notable.

There are no clear clusters of locations. Two locations with the highest

level of interest in immigration are New Philadelphia and Argyroupouli

(Greece) and the three locations with the lowest levels of interest are

Rimavska Sobota and Trnava (Slovakia) and Ozd (Hungary).

Though most countries had a similar mean score between the two

locations, there were clear clusters of countries. Finland and Estonia had

the largest differences, respectively 1.31 and 1.23. Next was the UK and

Russia, which had differences of 0.86 and 0.80. The third duo consisted of

Hungary and Georgia with 0.49 and 0.41 respectively.

MYPLACE August 2014


Q5_2: Interest in LGBT rights (lesbian, gay, bisexual, and transgender)


The variable has a minimum value 0, indicating no interest in LGBT

rights (lesbian, gay, bisexual, and transgender) and a maximum value

10, indicating maximum possible interest in LGBT rights.


Graph 4.5c: Interest in LGBT rights and issues by location (mean

values)

Narrative

The overall mean of the index is 3.70 and the standard deviation is

3.13. This varies by location, with the highest mean score at Sant

Cugat (Spain) with 6.11, and the lowest mean score at Narva (Estonia)

with 1.05. The differences across sites are quite notable.

There are no clear clusters of locations. Two locations with the highest

level of interest in LGBT rights are Sant Cugat (Spain) and Rostock in

eastern Germany and the two locations with the lowest levels of

interest are Narva (Estonia) and Forstate & Jaunbuve (Latvia).

Estonian and Finland had the largest differences between locations,

respectively 2.27 and 1.79. The next five countries were Latvia (1.10),

Spain (1.01), UK (0.92), Greece (0.82) and Denmark (0.81).

MYPLACE August 2014

MYPLACE: FP7-266831 www.fp7-myplace.eu Deliverable 4.6 Europe wide thematic report -DRAFT of 431

4.6 Closeness of political views with family and peers (Q20)

Political socialisation is a process consisting of the acquisition of political values, principles

and habits that takes place in adolescence and the first stages of youth. Usually this process

is based on the transmission of political dispositions from one generation to the next. The

family plays a central role in the socialisation process, as it is the main agent in the

generational transmission of political values and views. This is the reason why family has

been identified by the classical studies on socialisation as the main socialisation agent

(Dawson and Prewitt 1969; Jaros 1973). However, some studies have pointed out that as

teenagers become older, the influence of family, even if it is still crucial, becomes weaker

and subsequently, peer influence becomes more and more important (Dostie-Goulet 2009).

Gordon and Taft (2011) consider that this peer influence is becoming more important,

particularly to politically active young people.

The MYPLACE survey contains a question that asks the respondent to place the closeness of

or the distance between their political views with different people on a scale of 0-10: the

father, the mother, the closest brother or sister, the closest grandparent, the boyfriend,

girlfriend or partner and the best friend. The question provides an opportunity to test the

influence of the environment on the construction of the political ideas and values.

Specifically we want to compare the effect of both family and peers on transmitting political

views.

To do so, we have created two variables indicating the declared closeness of political views

with family and with peers. For the family, the closeness indicator that we have taken into

account is the closeness with father, mother and the closest grandparent. We have left out

the closeness of political views with brothers and sisters because we have considered that

they can be in a middle way from family to peers. Brothers and sisters are obviously family

members but they potentially belong to the same generation as the respondent. In this

case, we want to contrast the role of family with the role of the peer group, so we preferred

to have a purer comparison. For the peers’ political views’ closeness, we have considered

the answers about the closeness with the best friend and with the partner (or

boyfriend/girlfriend). In order to minimise the missing responses, when the respondent

answers that he or she does not have a partner, the indicator considers only the best friend

answer. In this same way, in the cases where there are no grandparents, the answer about

the closeness to the mother’s political views is taken.

MYPLACE August 2014


Q20: Closeness of political views Variable Description The graph contains the results of two variables: closeness of political views with family and with peers. Both are on a 0-10 scale where 0 is very distant and 10 is very close. Both variables have good scale reliability: Family closeness is constructed through father, mother and grandparents’ closeness questions with a Cronbach Alpha 0.82. The peer’s variable captures the closeness of political views with partner (or boyfriend/girlfriend) and with best friend that has a significant Pearson correlation 0.612.

Graph 4.6: Closeness of political views with family and peers

Narrative Locations in the graph have been ordered by family’s closeness of political views (orange bar). The first aspect that emerges from the data is that there is not a great variation across locations on the influence of family and peers on respondents’ political views. For the case of family, the level of closeness goes from a 5.39 in Kupchino (Russia) to a 7.34 in Sopron (Hungary). For the case of peers, from a 5.8 in Coventry (UK) to a 7.60 in Sopron (Hungary). Hungarian cities show the highest level of closeness of political views both for family and peers. In the case of family, Spanish cities are the next. On the other side, Danish cities are where young people declare themselves to be more distant from both family and peers’ political values and ideas. However, there is a part of this coincidence in the extreme positions where we cannot identify a clear country pattern in any of the two variables. It is also interesting to note that in all 30 locations the declared closeness to peers is always slightly higher than the closeness to family, apart from Sant Cugat (Spain) and Barrerio (Portugal). However, in none of these locations is the difference remarkable..

MYPLACE August 2014

MYPLACE: FP7-266831 www.fp7-myplace.eu Deliverable 4.6 Europe wide thematic report

of 431

4.7 Multi-level Modelling

4.7.1 Model 1: Political Interest

The purpose of multilevel modelling is to explain variation of political interest in terms of

individual and environmental factors.

For the purposes of this analysis, a cumulative index based on Q1 was developed. The main

reasons for selecting the Q1 are presented below:

The items are related to mental predispositions, which is what most of the authors in

the literature review considered to constitute the core of political interest.

The content of the question allows interpretation that it measures respondents’

general interest towards society, public affairs and politics.

Responses to the five items have no internal, hidden structure, suggesting that a

single cumulative index is an appropriate measure to be used.

The reliability measure of the index is sufficiently high.

Dependent variable

Our dependent variable is ‘political interest’, measured in the MYPLACE survey by the five

items question Q1 ‘How interested would you say you are in issues to do with …’ The scales

were reversed and as a result, values of the cumulative scale stretch from 0 (‘no interest in

politics) to 15 (‘maximum interest in politics’). To meet the needs of multilevel analysis, only

those respondents were selected which had valid values on all independent variables. The

number of such cases was 12,803 (75.6% of the total sample). The distribution of the index

is not normal, but is sufficiently close to normal distribution to undertake these analyses.

Multilevel modelling

The first question which needs to be answered before proceeding to multilevel modelling,

asks is the multilevel model plausible in the first place? Using a multilevel model is plausible

when the amount of variation at contextual level is significant. A common threshold is 10%.

If it is less than 10%, then a multilevel model would add little or nothing to our

understanding of what explains variation of the dependent variable.

Empty model

The first step in multilevel modelling is to estimate the amount of variation at contextual

level. For that, analysis of the empty model is conducted. Analysis of the empty model

shows that the share of variance at the level of research sites, accounts for 10% of variance

MYPLACE August 2014


of 431

in the political interest index Q1. A 10% threshold is generally considered as a lower band

when deciding whether to use a multilevel model. Thus, this case of analysis of political

interest index is lying exactly on the borderline where multilevel modelling loses its salience.

In spite of that, multilevel analysis was carried out.

Random intercept model with individual level explanatory variables

The list of individual level explanatory variables in this model includes the following items:

1. The level of education, Q60_isced (recoded into a 4 category variable)

2. Gender Q54

3. Age. Age1

4. Social class class3

5. Majority/minority status Identity

6. Employment status Q61R (as a set of dummy variables, reference category enrolled

in education)

7. Political knowledge Q28s

8. Mother’s voting activism, respondent assessed Q17_2

9. Father’s voting activism, respondent assessed Q17_1

10. Mother’s interest in politics, respondent assessed Q2_3

11. Mother’s interest in politics, respondent assessed Q2_2

Random intercept model with individual and contextual variables

On the next stage, a contextual level variable was added: a variable indicating whether

elections did take place in 2012 (yes = 1, no = 0). Since elections could be an event which

potentially arouse interest in politics, it can be expected to see a higher level of political

interest in countries where elections in 2012 did take place and a lower level of interest

where elections were not held.

Results

Below are the three models compared, to assess which set of variables is appropriate for

the model.

Table 4.7.1a Variance components of regression models explaining political interest

Empty

model

(random

intercept

only)

With

individual

level

explanatory

var’s

With

country

level

explanatory

var’s

σ (individual

level) 8.145 7.212 7.212

Political

interest

σ (country

level) 0.901 .818 0.819

MYPLACE August 2014


of 431

-loglikelihood 63302.510 61794.553 61793.905

ML estimates 63300.848 61730.389 61729.353

Changes in loglikelihood

The difference in loglikelihood between the empty model and the model with individual

level variables only is 1570.459 (63300.848-61730.3897). This is a highly significant

difference which means that the model using individual level variables is better that the

empty model. The difference in loglikelihood between the model with individual level

variables only and the full model, containing also one country level variable, is 1.036

(61730.389-61729.353). This change is not statistically significant which means that the

model containing contextual level model is not significantly better than the model

containing individual level variables only.

Comparison of the three models leads to the conclusion that the model containing only

individual level variables describes variation of the political interest index sufficiently well.

Including a variable, which indicated whether elections had taken place or not was

motivated by the wish to include a variable which would describe political environment. It

can be assumed that during an election period, people become targets of a relatively

intense flow of political messages. This in turn has a potential to arouse their interest in

politics. However, the hypothesis was not confirmed – this did not happen, the level of

interest was not linked to elections. Evidently, elections are not perceived as a sufficiently

outstanding political event to increase significantly young people’s interest in politics.

Indeed, regularly occurring elections are far less significant events than the fall of Berlin Wall

or the major crises in Greece, which have been reported to increase people’s interest in

politics and public matters. This finding is well in line with earlier research where political

interest has been found to be rather stable, an “unmovable mover”.

We can conclude that interest in politics is to a large extent determined by the following

individual level variables:

Level of education: higher educational attainment means also higher political

interest.

Age: higher age means also higher political interest.

Majority/minority group membership: belonging to the minority group means also

higher political interest.

Political knowledge: higher score on the political knowledge scale means also higher

political interest.

Mother’s frequency of voting: higher frequency means also higher political interest.

Father’s political interest: higher interest means also higher political interest.

7 The estimates were computed using maximum likelihood method for estimation parameter values, not the

restricted maximum likelihood method.

MYPLACE August 2014


of 431

Mother’s political interest: higher interest means also higher political interest.

The findings from the MYPLACE data corroborate earlier research findings about political

interest.

Interest in politics is positively linked with the level of education. Education is believed to

provide an individual with the mental tools and knowledge that are necessary for making

sense of political situation and developments and locate oneself in relation to them. This

line of reasoning is confirmed also by the finding that people scoring higher on the political

knowledge index also scored higher on political interest index. Even if we might experience

here a challenge in affirming that causal direction of the influence is from political

knowledge to political interest, when we take into account both level of education (which

cannot be changed by political interest) and political knowledge, then we can safely

conclude that education influences political interest and that the relationship is positive.

Age too has positive link to political interest. This, however, is most likely to a large extent

caused by education – in this age group, higher education and higher age are positively

correlated.

The presence of parents’ influence is in line with earlier research whereby the influence of

family has been found to be positive and also of crucial importance. However, since the

level of parents’ political interest is respondent-assessed, the validity of this measure is hard

to ascertain. It would be natural when a young person whose level of political interest is

high, talks to his or her parents quite frequently about politics and public affairs. From that,

he or she might conclude that also his or her parents have a high level of political interest.

Therefore ideally would it require an independent measurement, which we do not have in

this research.

Higher political interest of the minority group needs further research. At the present stage,

it can be explained with a reference to a weaker political position in the society, which

causes increased interest in politics and public affairs to be informed about various

developments. The developments could be seen either as opportunities, which could be

seized and utilised by members of a minority group, or as hazards, which need be avoided.

However, at the present stage this explanation remains in the status of hypothesis.

Table 4.7.1b Political Interest: Estimates of Fixed Effects (restricted maximum likelihood

estimates)

Parameter Estimate Std. Error T Sig.

Intercept 8.771 0.356 24.657 0.000

Level of education 0.138 0.027 5.035 0.000

MYPLACE August 2014


of 431

Gender -0.068 0.048 -1.393 0.164

Age (years) 0.063 0.012 5.347 0.000

social class derived from parents’ occupation and

education -0.007 0.022 -0.322 0.748

Majority/minority status in society (0=minority,

1=majority) -0.354 0.076 -4.668 0.000

Activity status: endrolled in education Ref.

Employed -0.050 0.066 -0.760 0.448

Unemployed -0.119 0.088 -1.350 0.177

Other -0.034 0.106 -0.324 0.746

Level of political knowledge 0.511 0.028 18.059 0.000

Father’s frequency of voting. respondent assessed -0.049 0.031 -1.575 0.115

Mother’s frequency of voting. respondent assessed -0.108 0.031 -3.481 0.001

Father’s interest in politics. respondent assessed -0.444 0.036 -12.488 0.000

Mother’s interest in politics. respondent assessed -0.420 0.035 -11.860 0.000

Elections. yes/no 0.335 0.337 0.992 0.330

4.7.2 Model 2: Closeness of Political Views

The descriptive data does not show a great difference between locations in the levels of

political views between young people and, their families and peers. However, to confirm

this, a multi-level model was developed. The table below shows that almost all the variance

in both variables can be explained by differences between individuals. The Intra-class

correlation indicates that only 3.1% of the variance of closeness of political views with

family can be attributed to differences across locations. For the case of peers’ closeness, the

results are equally clear, a 3.6% variance.

Table 4.7.2a Variance components of regression models explaining closeness of political views with family and peers

Empty model

(random intercept

only)

With individual

level explanatory

var’s Closeness of political views with family

σ (individual level) 5.908 5.806 σ (location level) 0.188 0.219 Intra-class correlation

0.031 0.036

Closeness of political views with peers

σ (individual level) 6.350 6.234 σ (location level) 0.237 0.269 Intra-class correlation

0.036 0.041

MYPLACE August 2014


of 431

This first analysis of the empty multi-level model suggests an explanatory model based on

individual variables. Basic socio-demographic variables are included, i.e. sex and age

(treated as a continuous variable (from 16 to 25). The model also contains a variable for the

activity status with three categories (Working, Studying and Unemployed or inactive) where

‘Working’ is the reference category. Level of education is included by a dichotomous

variable indicating young people with a post-secondary degree. The social origin of the

individual is captured by a continuous variable that takes information about the labour

situation and the level of education of parents, where a higher score in the variable

indicates a higher social position of parents. Finally, we have introduced two variables on

the political involvement of the respondent: The first is probably the most used variable in

political attitudes: Interest in politics (a dichotomous variable). The second is a 0-60 scale

capturing the level of political activism of the individual from a battery of 20 different

political actions that can be undertaken once, twice or three or more times in the last 12

months. The following table shows the estimates of these variables in a regression model

for closeness of political views with family and with peers.

Table 4.7.2b Parameter estimates of regression models explaining Closeness of political views with family (n1=13.671, n2=30) and peers (n1=13.124, n2=30)

Closeness of political views with family

Closeness of political views with peers

SE SE

Constant 5.214 *** 0.232 5.542 *** 0.247

Individual level variables

Activity status: Studying 0.017 0.056 0.093 0.059

Activity status: Unemployed or inactive

-0.111 0.070 0.041 0.074

Activity status: Working (ref. cat.) Ref. 0.000 Ref.

Age 0.015 ~ 0.009 0.037 *** 0.010

Sex (female) 0.174 *** 0.042 0.063 0.044

Post-Secondary education 0.057 0.060 0.035 0.063

Social Class of origin 0.402 *** 0.063 0.113 ~ 0.066

Interest in politics 0.364 *** 0.045 0.092 ~ 0.048

Non-electoral participation -0.016 *** 0.004 0.011 ** 0.004 ***p<0.001, **p<0.01, *p<0.05, ~p<0.1, two-tailed tests For both dependent variables, activity status and the level of education do not appear to

have an effect. The results show that age has a positive effect on the declared closeness of

political views both with family and peers. Thus, as teenagers get older, they tend to

strengthen their political ideas and values with their environment. However, we can point

out that this effect is stronger in the case of peers. This result would confirm the idea that

the effect of peers on youth political socialisation rises as young people gain independence

MYPLACE August 2014


of 431

from their family (Dostie-Goulet, 2009). Also in both cases, social class of origin and interest

in politics have a positive effect on political closeness with family and peers. Here, the effect

is stronger for family political views’ closeness. Regarding the effect of social class, it is

possible that there is more political discussion in families with higher levels of education and

that, political views are more defined, thus playing a positive role in transmitting political

ideas from one generation to another.

There are also a couple of variables that have a different impact on the closeness of political

views with family and with peers. Being a woman, for example, has a strong positive effect

on sharing the family’s political views but it does not have any effect for the case of peers.

The other variable with different effects on the closeness of family or peers is the level of

political activism of the respondent (non-electoral participation variable). This has a

negative effect on the sharing of political values and ideas with family, which means that

young people that are more politically active tend to feel more distant from their family’s

political ideas compared to more passive young people. By contrast, the level of political

activism has a positive effect on the closeness of political views with peers. These results

from the MYPLACE survey seem to confirm what Gordon and Taft (2011) report from a

qualitative analysis, in young activists it is the role of the peer-group on political socialisation

that is crucial. With these results we can confirm that young activists share the political

views of their peers more than non-activists and that they are more independent from their

family’s political ideas than young people who are politically inactive.

4.8 Summary

Following the dominant approach in the literature, which focuses on attitudinal as opposed

to behavioural aspects of political interest, the following definition of political interest was

adopted here: political interest is ‘curiosity, relation, attention and vigilance to political

issues and public affairs‘ (Schmid, 2004).

Political interest has been researched mainly within the framework of political activism. It

has been analysed as a factor, which influences a range of political attitudes, opinions and

forms of political activism. Political interest as a dependent variable, as a phenomenon in its

own right, has received relatively little research attention. There is no thorough

understanding of how it forms and what determines its levels. However, it is clear that the

level of political interest is very stable over the individual lifecourse and that, it forms to a

significant extent in quite early years, being strongly influenced by one’s parents and family.

Later in life, it fluctuates only as a result of very significant social disturbances and changes

such as the fall of Berlin Wall and the collapse of the Eastern Bloc in the late 1980s. Less

MYPLACE August 2014


of 431

significant social and political events do not have sufficient weight to alter its levels

significantly.

MYPLACE data showed that the political interest of young people is divided. Across all

countries, more people reported that they were ‘not interested’ in politics than those who

reported that they were ‘interested’ in politics (58%:42%). The level varied across countries;

young people were interested in politics in some locations (in eastern Germany, Greece and

Spain) while, in other countries, they were not very interested (Croatia, Latvia, Finland and

Estonia). Youth was most interested in politics related to the country they lived in and the

municipality, they lived in. Interest was lowest towards countries neighbouring their

country. Europe and the immediate surroundings were of medium interest.

Employment, housing and environment were the three issues that were of interest for the

highest percentage of young people. The European Union, immigration and LGBT rights, in

contrast, scored lowest on the scale of political interest. Television and the Internet were

the two information channels used by young people to obtain information about politics

and public affairs. Radio and newspapers were far less popular. Levels of different aspects of

political interest vary across sites, but differences usually are not large. When sites are

ordered from highest to lowest then the differences between immediate neighbours are

incremental. In the case of most questions, no single research locations are characterised

by exceptionally high or low values of political interest. Neither are there concrete clusters

of sites.

Across all five questions on political interest, certain sites tend to report systematically

lower or higher values. The pattern is far from deterministic and contains exceptions but

one can distinguish between sites where political interest tends to be higher and where it

tends to be lower. To give a very ‘broad brush’ ranking of sites, a cumulative scale summing

responses from the entire five question block was developed and sites were ordered

according to mean value of the index. The following 10 sites tended to have relatively high

political interest:

1. Coventry (the highest average political interest)

2. Odense East

3. Jena

4. Odense Center

5. Vic

6. Bremen

7. Sant Cugat del Vallès

8. Lumiar

9. Rostock

10. Telavi

MYPLACE August 2014


of 431

Among the sites where political interest tends to be low, are:

1. Pescenica Zitnjak

2. Lieksa and Nurmes

3. Nuneaton

4. Narva

5. Vyborg

6. Ozd

7. Trnava

8. Rimavska Sobota

9. Sopron

10. Forstate and Jaunbuve (the lowest average political interest)

This ordering of sites was obtained from summarising responses to all items in the five

questions. This index was not used earlier and thus the ranking of sites might vary from that

presented above. When using separate questions (Q1 to Q5) for ordering the sites, there is

some variation in rankings of the places. However, sites that scored high on one of the

questions, and appeared in the group of sites with high political interest, did not score so

low on any other question such that it also appeared in the group of sites with low political

interest. The opposite was also true. To conclude, despite some variation, the sites tended

to group in categories of low and high political interest. Detailed information is presented in

the graphs and tables above.

For multilevel analysis, the cumulative index built of responses to the first question (interest

in five different administrative/political/geographical units) was used. The level of political

interest was to a large extent explained by individual level variables (90%) and the role of

localities is small (10%). The variables, which significantly co-vary with the index of political

interest, are the following:


interest.


Majority/minority group membership: belonging to a minority group means also



political interest.





interest. Interest in politics is positively associated with the level of education. Education is

MYPLACE August 2014


of 431

believed to provide an individual with the skills and knowledge that are necessary for

making sense of political situations and developments and locating oneself in relation to

them. This line of reasoning is confirmed also by the finding that people scoring higher on

the political knowledge index also scored higher on the political interest index. Age too has a

positive association with political interest. This, however, is likely to a large extent to be a

product of education – in this age group, education and age are positively correlated. The

presence of parents’ influence is in line with earlier research, which indicates the influence

of family to be of crucial importance. However, since the level of parents’ political interest

is respondent-assessed, the validity of this measure is hard to ascertain. Reasons for higher

political interest among minority groups need further research. At the present stage, it can

be hypothesised that minorities are somewhat more interested in politics and public affairs.

In this section, we analysed the level of coincidence between the political ideas of young

people and their families and peers. Through this, we can gain an idea of the role of the

environment in the political socialisation process. By measuring the declared closeness of

political views with family and peers, it was shown that, while differences are not very high,

young people tend to feel closer to their peers than their family.

The MYPLACE data set – drawing on 30 different European locations – provides the

opportunity to analyse the role of context in shaping different phenomena. In this case,

however,it seems that context does not have an important role to play in explaining how

close or distant the political ideas and values of young people are from their families and

peers.

The analysis of the effect of different variables in defining the closeness of political ideas to

those of family and peers has shown interesting results concerning the role of political

activism. As young people participate more in political actions, they tend to share more

political views with their peers and less with their families. Thus, it seems that political

engagement facilitates the establishment of new political preferences, which are

independent of family background.

References

Hadjar, A. and Becker, R. (2006). Politisches Interesse und politische Partizipation. In:

Hadjar, A.; Becker R. (eds.): Die Bildungsexpansion. Erwartete und unerwartete Folgen.

Wiesbaden: VS Verlag für Sozialwissenschaften, p. 179-204.

MYPLACE August 2014


of 431

Holleque, M. (2011) Rethinking the Stability of Political Interest. Paper prepared for a

meeting of the University of Wisconsin’s Political Behavior Research Group, April 12th, 2011.

Kaase, M. (1990) Mass participation. In Jennings, M,K., van Deth, J.W. (eds.) Continuities in

political action, Walter de Gruyter, Berlin, New York.

Kaase, M., and Marsh, A. (1979) Political Action: A theoretical perspective. In Barnes, S.H.,

Kaase, M. (eds.) Political Action: Mass Participation in Five Western Democracies, Beverly

Hills, CA, Sage.

Kenski, K. and Stroud, N.J. (2006) Connections Between Internet Use and Political Efficacy,

Knowledge, and Participation. Journal of Broadcasting and Electronic Media 50(2), 2006, pp.

173–192.

Kuhn, H. P. and Schmid, C. (2004) Politisches Interesse, Mediennutzung und

Geschlechterdifferenz. Zwei Thesen zur Erklärung von Geschlechtsunterschieden im

politischen Interesse von Jugendlichen. In: Hoffmann, D.; Merkens, M. (eds.):

Jugendsoziologische Sozialisationstheorie. Impulse für die Jugendforschung, Weinheim and

München: Juventa, pp. 71-89.

Lange, D., Onken, H. and Slopinski, A. (2013) Politisches Interesse und Politische Bildung.

Zum Stand des Bürgerbewusstseins Jugendlicher und junger Erwachsener. Wiesbaden:

Springer VS.

Niedermayer, O. (2005) Bürger und Politik. Politische Orientierungen und Verhaltensweisen

der Deutschen. 2. Updated edition. Wiesbaden: VS Verlag.

Prior, M. (2010) You’ve Either Got It or You Don’t? The Stability of Political Interest over the

Life Cycle, The Journal of Politics, Vol. 72, No. 3, July 2010, Pp. 747–766.

Pritzker, S., Springer, M.J., McBride, A.M. (2012) Learning to Vote: Informing Political

Participation among College Students. CSD Working Papers No. 12–46.

Niemi R. G., Craig, S. C. and Mattei, F. (1991). Measuring Internal Political Efficacy in the

1988 National Election Study, The American Political Science Review. 85 (4): pp. 1407-1413.

Schmid, C. (2004) Politisches Interesse von Jugendlichen Eine Längsschnittuntersuchung zum

Einfluss von Eltern, Gleichaltrigen, Massenmedien und Schulunterricht. Wiesbaden:

Deutscher-Universitätsverlag.

Shani, D. (2009) On The Origins of Political Interest. A Dissertation Presented to the Faculty

of Princeton University in Candidacy for the Degree of Doctor of Philosophy. Recommended

for Acceptance by the Department of Politics. UMI Number: 3378025

Syal, R. (2012) What Are the Effects of Educational Mobility on Political Interest and

Participation in the Indian Electorate? Asian Survey, 52(2) (March/April 2012): pp. 423-439

Tenscher, J.; Scherer, P. (2012) Jugend, Politik und Medien. Politische Orientierungen und

Verhaltensweisen von Jugendlichen in Rheinland-Pfalz. Wien, Berlin: LIT Verlag.

MYPLACE August 2014


of 431

Utter, G.H. (2011) Youth and political participation: a reference handbook. California , ABC-

CLIO, LLC,.

Van Deth, J.W. (1990) Interest in politics. In Jennings, M,K., van Deth, J.W. (eds.)

Continuities in political action, Walter de Gruyter, Berlin, New York.

Verba, S., Schlozman, K.L. and Brady, H.E. (1995). Voice and Equality: Civic Voluntarism in

American Politics, Boston, Harvard University Press.

MYPLACE August 2014


of 431

Chapter 5: Attitudes and Trust

The country specific analyses (MYPLACE Deliverable 4.5.) has already showed that youth in

general have negative attitudes towards politicians and politics; have a tendency to believe

that politicians are corrupt, likely to be influenced by the rich and are not interested in

young people. For example in eastern Germany over 60% of youth perceive politicians as

corrupt, while in Croatia around 75% of youth perceive the same; about 60% of youth in

Spain considers that politicians are not interested in young people, while in Croatia only

about a tenth perceive politicians as interested in young people.

Additionally in country specific analyses several socio-demographic characteristics of youth

were confirmed as significant sources of differences in youth attitudes towards politicians

and politics such as age (e.g. Croatia, Georgia), level of education (e.g. Croatia, Georgia),

employment status (Georgia) and religious affiliation (in Jena in eastern Germany).

However, some country reports (e.g. eastern Germany) stressed no established significant

differences in youth attitudes towards politicians and politics with regard to those

characteristics (age, level of education, occupational status and social class), while some

reports stressed some additional correlates of attitudes towards politicians and politics such

as voting preferences (in Georgia).

Besides differences in attitudes towards politics and politicians on individual level socio-

demographic characteristics, some national reports (e.g. Croatia, Latvia) stressed the

difference with regard to locations (Croatia, Latvia, whereas attitudes were more negative in

economically poorer locations).

Thus generally speaking, findings to date, in addition to generally negative attitudes of

youth towards politicians and politics demonstrated that youth attitudes are determined by

individual characteristics, and in some cases by locations. However, also indicated was the

possibility that relationship between socio-demographic characteristic and attitudes could

depend on locations and countries as well as country level characteristics as possible

determinates of those attitudes.

Regarding trust in institutions, country level analyses indicate low to medium level of trust

among youth, especially towards national political institutions. For example, head of the

government and political parties were among lowest-ranked institutions in terms of trust in

Croatia and Spain. However, in some countries religious institutions (e.g. eastern Germany)

and banks (eastern Germany, western Germany) were lowest ranked. Some of the

institutions which were most trustworthy for young people are judiciary institutions, such as

courts and the police (in western Germany), or the army (UK followed by police and courts,

MYPLACE August 2014


of 431

Croatia), international institutions Greenpeace or Amnesty International (Spain, Croatia,

after the army).

Additionally, country specific analyses indicated some of the socio-demographic

characteristics as possible determinants of young people’s trust in institutions: gender (in

Croatia), age (eastern Germany, western Germany), and social class (Croatia, eastern

Germany) level of education (eastern Germany). Generally speaking, females, youth in the

higher class and with higher the educational attainment expressed a higher level of trust in

national, as well as in international institutions. However, established differences with

regard to individual socio-demographic characteristics were small. Besides socio-

demographic characteristics some country analyses tested some attitudinal individual

variables as determinants of institutional trust such as satisfaction with democracy

(Georgia), political knowledge (eastern Germany), confirming these variables as positive

predictor of trust.

Besides this, some teams reported difference in some of the indicators of the level of trust

between their two research locations, whereas lower level of trust was more characteristic

for generally speaking, poorer location (Peščenica in Croatia, the Pielinen Karelia area in

Finland, Lumiar in Portugal, and Nuneaton in UK).

Thus, similarly to the country specific findings relating to attitudes towards politicians and

politics, the cross country analyses indicates that countries (and in some cases locations

within the countries) are significant determinants of young people’s trust toward

institutions. Moreover, findings indicate that the effects of some socio-demographic

variables could depend on context (country or location), in addition, some additional

individual level variables could be an important source of difference in the youth’s level of

trust towards political and social institutions.

In this chapter, we present findings of the cross-country analyses based on the merged data

set from all locations. In the following sections we firstly describe youth attitudes towards

politicians and politics, as well as youth trust in institutions with regard to locations and

countries in more detail (together with details on the construction and testing of composite

variables based on the cross-country merged data set). After these analyses, we report the

findings of two multivariate analyses; first with trust in national parliament as the

dependent variable and second with trust in the European Commission as the dependent

variable.

In this part, we will test the following hypotheses, which we developed for this study:

• Youth trust in national Parliament as well trust towards the European Commission

are differentiated by individual level characteristics as well as location and country level

characteristics

MYPLACE August 2014


of 431

• Relationship between individual level determinants of trust towards national

parliament and the European Commission depends on context (specific location or country).

Literature

The existing research consistently shows that youth, with regard to their attitudes to politics

and politicians, can be characterised as dissatisfied and alienated from the political process

(Henn and Foard, 2012; White et al.,2000). For example, data from the international

CivicWeb project showed that youth in general saw politicians “as corrupt, boring, or hard

to understand; working only for their own interests; and far removed from the everyday

needs and realities of common citizens (Banaji and Buckingham, 2010:17). Moreover, such

negative youth attitudes could be, regarded as one of the sources of youth political

disengagement (at least from formal politics). However, as Marien argues “… critical

attitude towards the politicians in office is generally considered to be normal and healthy

part of democracy…However citizens should be able to put (some) trust in democratic

procedures and institutions (Marien, 2011:13).

Political trust is, usually defined as ‘a summary judgment that the system is responsive and

will do what is right even in the absence of constant scrutiny’ (Miller and Listhaug, 1990:

358). Thus, as Hooghe et al. (2014) stress, political trust can be considered as one of the

most important resources for a democratic political system. Research studies confirm the

important role of political trust in different citizens behaviour such as paying taxes (Orviska

and Hudson, 2003), policy preferences (Hetherington and Husser, 2012) as well voting

turnout (Jones and Hudson, 2000). In the context of considering the relation between

political trust and political participation or civic engagement in general it should be, noted

that there are many conflicting findings, which suggest that this relationship involves

complex interactions and contingencies (Levi and Stoker, 2000). For example, a recent study

(Hooghe and Marien, 2013) showed that political trust is positively associated with

institutional participation, while it is negatively associated with non – institutional

participation.

Data from international surveys e.g. European Social Survey shows a consistent pattern of

trust in different institutions: trust in implementing institutions is higher than in

representational institutions, while political parties and politicians receive relatively lowest

trust (Marien, 2011). Moreover, Mishler and Rose (2001) reported that citizens in post-

Communist societies express greatest distrust for political institutions, especially

parliaments and parties, while the military (as the least democratic institution) enjoys the

highest level of popular trust. Additionally, international institutions, especially in new

democracies, enjoy more trust than national institutions whereas level of citizens’ trust in

MYPLACE August 2014


of 431

European parliament is usually higher than level of trust in national parliament. One of the

explanations is that this higher level of trust in international institution is a consequence of a

lack of direct experience with these institutions (Marien, 2011). Marien (2011) augmented

this claim with the fact that number of missing values is usually higher for international

institutions than for national institutions.

However, there are consistent cross-country differences in the level of trust in institutions.

Whereas Nordic countries and Netherlands are, characterized by higher political trust,

Central and Eastern Europe countries are characterised by rather low levels of trust (Van der

Meer, 2010). Hakhverdian and Mayne (2012) distinguished three groups of countries based

on levels of trust in national political institutions (from round 4 of European Social Survey);

the first group of ‘highly trusting’ countries consists of Sweden, the Netherlands,

Switzerland, Norway, Finland, and Denmark. A second group of ‘moderately trusting’

countries consists of Slovenia, Estonia, Spain, France, Great Britain, Belgium, and Germany,

while a third group of ‘low trusting’ countries consists of Latvia, Hungary, Croatia, Poland,

the Czech Republic, Portugal, Romania, and Greece. Analyses by Hooghe and Wilkenfeld

(2008) showed that the country patterns of political trust among youth correspond to those

established among general samples: on the level of country aggregates, higher level of trust

was characteristic for youth in Scandinavian countries and lower for youth in Southern and

Eastern Europe countries. Recent analyses by Harteveld et al.(2013) showed a reverse

pattern for trust in EU institutions. According to Eurobaremeter data, the most EU-trusting

citizens are characteristic for the new, post-2004 member states and, to a lesser extent, the

Benelux countries. While in post-communist developing democracies where national trust is

low, evaluation of EU competence as well as trust in EU are high (Harteveld et al.,2013). A

recent study by Armingeon and Ceka (2014) demonstrated a significant drop in the level of

trust in EU between 2007 and 2011 during the economic crisis.

In conceptualising political trust, two main theoretical approaches have been distinguished

(Mishler and Rose, 2001; Hakhverdian and Mayne 2011). According to cultural theories,

political trust is a consequences of attitudes and values that are, learned early in life and are

transmitted from generation to generation (Inglehart, 1997; Putnam, 2000). Within this

approach, one of the main proximal determinants of political trust is interpersonal or dyadic

trust. On the other side, according to the institutional theories political trust depends on

institutional performance and citizens evaluations of performance (Hakhverdian and Mayne

2011). In accordance with this approach Van der Meer and Dekker (2011) conceptualized

trust in national institutions as a citizen’s rational evaluation of the relationship between

citizen and state, which depends on four aspects: competence, care, accountability and

reliability.

Both general approaches make the distinction between macro and micro level determinants

of trust. Cultural theories emphasise the national traditions as macro level determinants

MYPLACE August 2014


of 431

and individual socialisation experiences as micro level determinants. At the same time more

institutional approaches emphasise the aggregate performance of institutions, such as

those which could be operationalised by indicators of economic growth and level of

corruption as macro level determinants, while individual evaluations of institutional

performance are contingent upon individual experiences and attitudes and are regarded as

micro level determinants (Mishler and Rose, 2001).

Although political trust measures are common in many international and national surveys,

there is no consensus in the literature, which, is the best way to operationalise and analyse

political trust. As Marien (2011) emphasises, institutional trust has been conceptualised and

studied as both a one-dimensional and multidimensional attitude. Thus, some of the

authors (e.g. Michler and Rose 2001; Marien, 2011) analyse trust in institutions as a whole,

while others advocate using measures of trust towards specific institutions. Among those

specific institutions relatively frequently investigated were national parliament (e.g. Van der

Meer, 2010; Van der Meer and Dekker (2011), whereas “trust in parliament may be a good

measure because confidence in institutions is about something deeper and more

fundamental than trust in politicians or in particular governments” (Newton, 2013:?), as

well as European institutions (Roth et al.,2013).

Besides, as Levi and Stoker (2000) observed, survey research traditionally examines

individual level correlates of political trust, while macro level correlates have been the focus

of historical and comparative case studies. However in recent years there are more studies

which have employed multilevel statistical analyses which enable investigating individual as

well as macro level determinates of political trust (Van der Meer 2010; Van der Meer and

Dekker (2011), as well as their interaction. For example, Hakhverdian and Mayne (2012)

revealed that education at the individual level is negatively related to institutional trust in

corrupt societies and positively related to institutional trust in societies not regarded as

corrupt.

Regarding trust in the national parliament Van der Meer (2010) confirmed by multilevel

modelling that corruption, the electoral system, and former regime (communist) type are

important determinants of the cross-national differences in trust, while economic

performance is not related to trust in parliament. Additionally, Van der Meer and Dekker

(2011) revealed a negative effect of a recent communist regime is the same for both young

and old, indicating that the perceived negative effects of communist rule on political trust

endures even after a decade since its downfall (van der Mere and Dekker, 2011:104).

Regarding trust in European institutions, Harteveld et al.(2013) conceptualised that trust in

the EU could be the result of three different processes; rational, based on evaluations about

the performance and procedures of the European Union; emotional, based on the salience

of a European identity as well national identity, and extrapolation, based on national trust

MYPLACE August 2014


of 431

and therefore unrelated to the European Union itself. Based on a multilevel analyses they

concluded that citizens’ trust in the EU can be predicted by their trust in national

institutions, regardless of their rational evaluation or emotional affiliation. Armingeon and

Ceka (2014) also confirmed that the trust in EU strongly depends on the trust in national

institutions, but also on individual evaluations of the national economy. However Munoz et

al.(2011) showed that the relationship between trust in national institutions and trust in

European institutions could be different at individual and country levels; at the country level

- living in a country with highly trusted and well-performing institutions hinders trust in the

European Parliament, while at the individual level those who are more trusting tend to be

so at both the national and the European level.

Although investigating the determinants of political trust is a relatively common research

topic, studies investigating political trust (trust in the EU especially) and their correlates

specifically among youth are relatively scarce (Hooghe and Wilkenfeld, 2008; Schoon and

Cheng, 2011; Torney-Purta et al.,2004). We believe there are no studies related to youth

political trust or youth trust towards specific national or European institutions which employ

multilevel analyses, investigating individual and country level determinates.

5.1 Attitudes towards politicians and politics (Q6)

Respondents were asked about their attitudes towards politics and politicians. The following

3 items with each having five points ranging from ‘strongly disagree’ to ‘strongly agree’ were

used to assess young people’s attitudes:

Q6_1: Politicians are interested in young people like me

Q6_2: Politicians are corrupt

Q6_3: The rich have too much influence over politics.

These three items are loaded under one factor which, explained 58.13% of their variance.

Further analysis of data suggests that these three items constitute a scale of acceptable

reliability α ≥ .60. Nevertheless, we have decided to use only two items (Q6_2 and Q6_3) for

a construction of ‘Attitude towards politicians and politics’ scale because in most of the

locations this solution has smaller number of missing cases and higher reliability as

presented in Table 5.1. However, it should be, noticed that in many locations the Cronbach

alpha is below acceptable level, for this option too. An ordinal by ordinal, Gamma

correlation of 0.588 suggests there is a strong relationship between the two variables and

theoretically they can be combined.

MYPLACE August 2014


of 431

Individual country Cronbach alpha values varied between .433 in Georgia and .732 in

Hungary, while individual country ordinal Gamma correlation varied between .341 in

Georgia and .704 in Hungary (all p’s <.001).

Table 5.1 Cronbach Alpha and percentage of missing cases per location for three and two

item ‘Attitude towards politicians and politics’ scale

Country Cronbach’s Alpha % of missing cases Cronbach’s Alpha Gamma correlation % of missing casesCroatia ,69 4,1% ,71 ,680 3,0%

Denmark ,43 3,6% ,52 ,411 3,4%

Estonia ,61 10,0% ,65 ,589 8,5%

Finland ,62 9,3% ,68 ,581 7,7%

Georgia ,26 18,1% ,43 ,341 16,3%

West

Germany,48 4,5% ,47 ,409 3,7%

East

Germany,46 2,5% ,50 ,425 2,1%

Greece ,70 4,3% ,59 ,609 3,8%

Hungary ,55 8,8% ,73 ,704 7,9%

Latvia ,55 4,0% ,63 ,577 3,3%

Portugal ,57 2,9% ,51 ,456 1,8%

Russia ,48 8,3% ,64 ,594 7,3%

Slovakia ,55 9,6% ,65 ,637 7,1%

Spain ,51 2,9% ,45 ,421 2,6%

UK ,54 2,2% ,57 ,496 2,1%

Total sample ,63 6,4% ,66 ,588 5,4%

Attitude towards politicians and politics

(3 items: Q6_1R + Q6_2_R+Q6_3R)

Attitude towards politicians and politics (2 items: Q6_2_R +

Q6_3R)

MYPLACE August 2014


of 431

Q6: Attitudes towards politicians and politics (Q6_2 & Q6_3)


An ordinal by ordinal Gamma correlation 0.588 suggest there is a strong

relationship between the two variables (Q6_2: Politicians are corrupt and

Q6_3: The rich have too much influence over politics) and theoretically they

can be combined. The scale, has a minimum of zero (negative attitudes

towards politicians and politics) to 8 (positive views towards politicians and

politics).

Graph 5.1a: Mean positive attitudes towards politicians and politics

by location

Narrative

The overall mean for attitudes towards politicians and politics is 2.46

(n=16,018, sd=1.72) out of a possible score of 8. This varies by location, with

the highest mean score at Odense Center (Denmark) with 4.55, and the

lowest mean score in New Philadelphia (Greece) with 1.14. However, it

should be, noted that youth in almost all locations on average has negative

attitudes towards politicians (except Denmark and Finland where average

attitude is neutral).

There are clear clusters of locations; with locations in Denmark, Finland,

Georgia and Estonia all in the top third, with neutral or less negative

attitudes towards politicians and politics. Locations in Croatia, Slovakia,

Portugal, Spain and Greece are, clustered in the bottom third, indicating

very negative attitude toward politicians on those locations.

Most Countries had similar mean scores between the two locations; notable

exceptions were found in Finland (Kuopio and Lieksa & Nurmes) with 0.5

difference between two mean scores.

MYPLACE August 2014


of 431

5.2 Trust in institutions and organisations (Q7)

Respondents’ trust in 13 different institutions and organisations was, assessed using a 0-10

scales. These 13 institutions and organisations are the courts, the police, the head of

government (PM), the media (national press and TV), banks, the United Nations, the

European Commission, national parliament, Amnesty International, the army, Greenpeace,

religious institutions and political parties.

Due to a high number of missing values on certain items (The United Nations – 1,147 (6.8%),

The European Commission – 1,602 (9.5%), Amnesty International 3,323 (19.6%) and

Greenpeace 2,088 (12.3%)), they were, excluded from the combined analysis. Factor

analysis conducted on the remaining nine items revealed one factor (eigenvalue greater

than 1, explaining 50.72 % variance), whereas all nine items are loaded under that factor.

This scale has a good reliability (α=0.872), and it could be regarded as good measure of

overall trust in institutions.

However, taking into account the main topics of this study (e.g. political engagement) as a

more relevant measure in context of this study we also present a descriptive analysis of

trust in core national political institutions (composite measure) as well as one item measure

of trust in national parliament and trust towards the European Commission.

A composite measure of trust in national political institutions is, based on three items;

trust towards head of government/PM (Q7_3), trust in national parliament (Q7_8) and trust

in political parties (Q7_13). A factor analysis of these three items showed that all of them

load under a single factor with excellent reliability (α = 0.85), while individual country

Cronbach alpha varied between 0.651 in Georgia and 0.913 in Hungary. The scale has a

minimum of zero (no trust in national political institutions at all) to 30 (complete trust in

national political institutions.)

Using single item measure of trust toward national parliament (Q7_8) as a dependent

variable is relatively common in political trust literature (e.g. Van der Meer, 2010) and

therefore we decided to present descriptive results for this item separately as well.

Additionally, “trust in parliament may be a good measure because confidence in institutions

is about something deeper and more fundamental than trust in politicians or in particular

governments” (Newton, 2013).

Trust toward the European Commission was, examined individually, given the importance

of the European perspective to the MYPLACE project. The approach of examining trust

towards European institutions is, often used in several surveys like Eurobarometer or

European Social Survey (Roth et al.,2013).

MYPLACE August 2014


of 431

Table 5.2 Percentage of Missing Values by question

Country

Trust in

national

parliament

Trust in

European

Commission

Croatia 1.5% 7.8%

Denmark 1.2% 13.3%

Estonia 2.5% 12.4%

Finland 4.9% 14.7%

Georgia 2.6% 12.6%

Western

Germany 4.3% 11.3%

Eastern

Germany 1.4% 5.9%

Greece 1.3% 5.9%

Hungary 5.1% 10.7%

Latvia 0.8% 4.5%

Portugal 1.5% 3.0%

Russia 4.8% 13.1%

Slovakia 3.3% 7.9%

Spain 4.4% 7.4%

UK 2.7% 14.8%

Total 2.8% 10.4%

MYPLACE August 2014


of 431

Q7: Overall trust in institutions and organisations scale (9 item scale) Graph 5.2a: Mean overall trust in institutions and organisations by

location


A factor analysis of these nine items showed that all items are loaded under

a single factor with good reliability (α = 0.872). We therefore constructed an

index composed of responses from each item and analysed this scale by

location. The scale, has a minimum of zero (no overall trust in institutions

and organisations) to 90 (complete trust in institutions and organisations).

Individual country Cronbach alpha varied between 0.786 in Denmark and

0.927 in Hungary.

Narrative

The overall mean for trust in institutions and organisations is 43.63

(n=15,357, sd=16.94). This varies by location, with the highest mean score at

Kuopio (Finland) with 58.66, and the lowest mean score at Argyroupouli

(Greece) with 30.97.

There are clear clusters of locations; with locations in Finland, Georgia,

Denmark and western Germany all in the top third, showing higher overall

trust in institutions and organisations. Greece, Croatia, Spain and Hungary

are, clustered in the bottom third.

Most countries had similar mean scores between the two locations; notable

exceptions were; Estonia (Narva area and Tartu) 9.78 difference, Russia

(Kupchino and Vyborg) 6.75 difference, and Finland (Lieksa & Nurmes and

Kuopio) 5.47 difference between two mean scores.

MYPLACE August 2014


of 431

Q7: Trust towards core national political institutions (Q7_3 head of

government PM, Q7_8 parliament and Q7_13 political parties)

Graph 5.2b: Trust towards core national political institutions by

location


A factor analysis of these three items showed that all of them load under a

single factor with excellent reliability (α = 0.85). We therefore constructed an

index composed of responses from each item and analysed this scale by

location. The scale has a minimum of zero (no trust in national political

institutions at all) to 30 (complete trust in national political institutions).

Individual country Cronbach alpha varied between 0.651 in Georgia and 0.913

in Hungary.

Narrative

The overall mean for trust in national political institutions is 12.54 (n=16,128,

sd=6.88) out of a possible 30. This varies by location, with the highest mean

score at Kuopio (Finland) with 18.54, and the lowest mean score at

Argyroupouli (Greece) with 7.27.

There are clear clusters of locations; with locations in Finland, Denmark and

Georgia all in the top third, stating high trust in national political institutions.

Locations in Greece, Croatia, Portugal and Spain are, clustered in the bottom

third stating less trust in national political institutions.

Most Countries had similar mean scores between the two locations; notable

exceptions were; Latvia (Agenskalns and Forstate & Jaunbuve) 3.28

difference, Estonia (Narva area and Tartu) 2.6 difference and Russia

(Kupchino and Vyborg), 2.56 difference.

MYPLACE August 2014


of 431

Q7_8: Trust in national parliament Graph 5.2c: Mean Trust in national parliament


The distribution of the responses are normally distributed and

provides a good eleven-point scale, where high values indicate higher

trust in national parliament.

Narrative

The overall mean for trust in national parliament is 4.50 (n=16,465,

sd=2.58). This varies by location, with the highest mean score at

Odense Center (Denmark) with 6.49, and the lowest mean score at

Pescenica (Croatia) with 2.61.

There are clear clusters of locations; with locations in Denmark,

Finland, Georgia and both eastern and western Germany all in the top

third, showing higher trust in national parliament. Croatia, Greece and

Hungary are, clustered in the bottom third.

Most Countries had similar mean score between the two locations;

notable exceptions were; Latvia (Agenskalns and Forstate & Jaunbuve)

1.03 difference, Russia (Kupchino and Vyborg) 1.02 difference, and

United Kingdom (Coventry and Nuneaton) 0.99 difference between

two mean scores.

MYPLACE August 2014


of 431

Q7_7: Trust in European Commission Graph 5.2d: Mean Trust in European Commission




trust in the European Commission.

Narrative

The overall mean for trust in European Commission is 5.17 (n=15,333,

sd=2.39). This varies by location, with the highest mean score at Telavi

(Georgia) with 6.60, and the lowest mean score at two locations in

Greece Argyropouli (3.45) and New Philadephia 3.45.

There are clear clusters of locations; with locations in Georgia,

Denmark, and eastern Germany all in the top third, showing higher

trust in European Commission. Greece, Croatia and Hungary are,

clustered in the bottom third.

Most countries had similar mean score between the two locations;

notable exceptions were; United Kingdom (Coventry and Nuneaton)

0.96 difference between two mean scores, Latvia (Agenskalns and

Forstate & Jaunbuve) with 0.86 difference and Estonia (Narva area

and Tartu) and Finland (Lieksa & Nurmes and Kuopio), both showing

0.80 difference between mean score trust in European Commission

between two locations.

MYPLACE August 2014

MYPLACE: FP7-266831www.fp7-myplace.eu Deliverable 4.6 Europe wide thematic report of 431


In this section, we present the results of multi-level modelling which took account of the

nested structure in our data set. First we present the effects of socio-demographic variables,

theoretically/empirically relevant individual level variables and contextual variables on the

trust in national parliament as a dependent variable (Model A), and latter we do the same

for trust in the European Commission as a dependent variable (Model B).

In order to establish the proportion of variability in our dependent variables associated with

individual level, location level and country level we ran a null (no predictors) model for both

of our dependent variables. Regarding trust in national parliament, 17.2% of the variance of

trust in parliament can be explained on the level of countries, while 2.4% of the variance can

be explained on the level of localities. Regarding trust in European commission, 7% of the

variance of trust in European commission can be explained on the level of countries, while

only 2.2% of the variance can be explained on the level of localities.

Due to a relatively low percentage of variance on the level of localities, we decided to

proceed with a two-level regression analysis with 16,935 individuals nested within 14

countries, ignoring the level of localities.

Model A: Trust in national parliament

Independent variables

Socio-demographic and individual level variables

After the null model, we present the results of a random intercept model with individual

level explanatory variables (Model 1). We included three socio-demographic variables - age,

gender and parental social class. Parental social class was used as a proxy for young

people’s social class because many young people in our sample still have not entered the

labour market. These socio-demographic variables were included as a control variables and

no specific hypothesis regarding their relationship with dependent variable were drawn.

In the next step of Model 1, we include six individual level variables. Two of them are more

general; social trust and life satisfaction. Social trust is expected to correlate with trust in

national parliament as Newton and Norris (2000) have already shown using World Value

Survey data. Both variables were measured by single item measures which is a common

approach used in many surveys (e.g. asking participants to assess on a 10 point scale

‘whether most people can be trusted, or that you can’t be too careful in dealing with

people’). The inclusion of a measure of life satisfaction is in line with recent evidence that

MYPLACE August 2014


linked psychological traits with citizens political attitudes and behaviour (Mondak and

Halperin, 2008). We have also decided to use several individual level variables more related

to politics as predictors of trust towards national parliament: individuals’ self-positioning on

a left-right ideology scale which is usually used as a predictor for political opinions on

national political issues and has been associated with trust in national institutions (King

1997), satisfaction with democracy (single item measure) and interest in politics. We used

a composite measure of interest in politics made out from Q1 (all items), Q2_1 and Q2_3 (all

items) as suggested in the scale and index construction (see Appendix A). Finally, we

hypothesise that the level of care our respondent’s perceive to get from politicians is related

to the amount of trust they will have in national parliament. Van der Meer and Dekker

(2011) described care as one of four aspects responsible for citizens’ trust in state. In this

analysis we used a proxy measure of care, operationalised through a single item measure of

perception of politicians interest in young people (“Politicians are interested in young

people like me”, Q6_1). All individual level variables were grand-mean centred.

Contextual variables

We included three contextual variables in the analysis. The first refers to welfare state types

proposed by Kaariainen and Lehtonen (2006). The remaining two contextual variables are

Corruption perception Index (CPI), a common measure in this field of research introduced

by Transparency International in 2002 which measures perceived level of corruption in

country’s public sector, and rule of law, a summation of indicators; which measure the

extent to which agents have confidence in and abide by the rules of society (Jan et al.,

2013). A brief description of these indices is included in appendix.

Results

The inclusion of individual level variables resulted in 28.6% decrease of the residual,

meaning that almost thirty percent of variance of trust in national parliament within a

certain country can be attributed to a set of individual level predictors included. Chi-square

test of deviances (loglikelihood = 24049.218; df = 9) confirms that this is a highly

significant difference. The inclusion of individual level variables led to 70.4% of reduction of

intercept variance, indicating the amount of the variance of trust in national parliament

between countries that can be attributed to include a set of individual level predictors.

In the next step we included contextual variables, which led to 39.4% reduction of intercept

variance compared to the model containing only individual level predictors. This difference

in the fit of the model in comparison to a model containing only individual level predictors is

significant at.05 level as indicated by chi-square test of deviance (loglikelihood = 7.81; df

= 3).

MYPLACE August 2014


Table 5.3.1 Variance components of regression models explaining trust in national

parliament

Empty model

(random intercept

only) – Model 0

With individual

level explanatory

var’s - Model 1

With country level

explanatory var’s -

Model 2

Trust in

European

commission

σ (individual level) 5.481 3.914 3.914

σ (country level) 1.235 0.366 0.222

-2 log likelihood 74820.501 50771.283 50763.865

df 3 12 15

Three socio-demographic predictors were included in the Model 1. Class of respondents and

respondents’ age were significant predictors of trust in national parliament (p<.001). In

comparison to the lowest class respondents, those from higher social classes indicate 0.203

increase in trust toward national parliament. Regarding age, younger respondents showed

more trust in national parliament. Gender was significant at .01 level, with women showing

more trust in national parliament. Among other individual level predictors, life satisfaction,

satisfaction with democracy, interest in politics and perceived lack of care were significant

predictors at .001 level, and the direction of these predictors was as expected. Respondents

who were more satisfied with democracy and life in general and those who have higher

interest in politics and believe that politicians are interested in young people showed higher

level of trust in national parliament. Also, lack of trust in politicians was associated with

lower trust in national parliament (p<.001).

In Model 2 we included three contextual variables out of which only corruption perception

was significant at .05 level indicating that countries with perceived lower corruption have

more trust in national parliament.

Van der Meer and Dekker (2011) conceptualised trust as citizens’ rational evaluation of

relationship between citizen and state, which depends on four aspects, namely,

competence, care, accountability and reliability. Care refers to citizens’ belief that object of

trust (in our case national parliament) will act in their best interest. Level of trust has been,

shown as an important country level variable that influence citizens distrust in political

system (Della Porta, 2000, as cited in van der Meer and Dekker, 2011). Since the individual

variable of care has shown to be a strong predictor of trust in parliament, in the next model

(Model 3) we were interested in establishing whether the slope of relationship between

perceived care and trust in national parliament varies across countries. Specifically, we are

interested in seeing if the slope varies across countries with different levels of corruption.

Because highly corrupted governments send clear messages that they do not care for their

citizens we hypothesise that more corrupted countries will enhance the care-trust slopes.

Therefore, this model requires the introduction of cross-level interactions between

MYPLACE August 2014


individual’s perception of politician’s interest in young people (care) and the countries’ level

of corruption (CPI). As a prerequisite for running this model we confirmed that slope

variance for relationship between care and trust in national parliament varies across

countries (Wald Z=2,115, p<.05). Chi-square test of deviance confirmed significant increase

in fit of this model in comparison to Model 2 (loglikelihood = 31.35; df = 2). However, as

evident from the last row in Table 5.3.2, the inclusion of this interaction term did not result

in statistical significance.

This difference in the fit of the model in comparison to a model containing only individual

level predictors is significant at .05 level as indicated by chi-square test of deviance

(loglikelihood = 7,81; df = 3).

Table 5.3.2 Determinants of trust in national parliament

Model 1 Model 2 Model 3

b p SE b p SE b p SE

Individual level predictors

Gender (1=M, 2=F) .109 ** .036 .109 ** .036 .109 ** .036

Class (0=lowest, 1=higher and high) .203 *** .038 .202 *** .038 .202 *** .038

Age -.021 *** .007 -.022 *** .007 -.022 *** .007

Left-right .017 * .008 .018 * .008 .018 * .008

Social trust -.012

.008 -.011

.008 -.008

.008

Life satisfaction .080 *** .009 .079 *** .009 .078 *** .009

Satisfaction with democracy .337 *** .009 .336 *** .009 .336 *** .009

Interest in politics .057 *** .004 .057 *** .004 .056 *** .004

Lack of care -.481 *** .019 -.480 *** .019 -.685 *** .126

Contextual predictors

Welfare state (0=other, 1=post-socialist)

-.309

.365 -.356

.371

Corruption perception index (CPI)

.048 * .019 .046 * .019

Rule of law

-.872

.494 -.863

.503

Lack of care * CPI

.003

.002

***p<0.001, **p<0.01, *p<0.05

Due to the large N we are taking into account only variables which are significant on p<0.01

MYPLACE August 2014


Model B: Trust in European Commission


Socio-demographic and individual level variables

Three socio-demographic variables - age, gender and parental social class - were included in

modelling of trust in European Commission. As in Model A, parental social class was, used as

a proxy for young people’s social class because many young people in our sample still have

not entered the labour market. There is evidence that women (Nelsen and Guth, 2000) and

older people (Arnold et al.,2012) have, more negative evaluations of, EU institutions, but no

specific hypothesis regarding relationship between these socio-demographic variables were

made.

In the next step of Model 1, we included five individual level variables. As in Model A, we

tested social trust and life satisfaction as possible individual level predictors of trust towards

European Commission. Social trust, measured with a common approach used in many

surveys (asking participants to assess on a 10 point scale whether most people can be

trusted, or that you can’t be too careful in dealing with people) is also expected to influence

trust in European Commission. Inclusion of measure of life satisfaction (single item

measure) is in line with recent evidence that linked psychological traits with citizens’

political attitudes and behaviour (Mondak and Halperin, 2008). Higher life satisfaction has,

been connected with higher trust in European institutions, including European Commission.

As Christine et al. (2012) noticed, European issues are becoming more incorporated in

domestic political discourse so we also included in the model individuals’ self-positioning on

a left-right ideology scale. This is usually used, as a predictor for political opinions on

national political issues, assuming that this variable might also have an effect on trust in

issues beyond national scope.

Satisfaction with the functioning of the national democratic system has also been

connected with trust in European institutions (Norris, 1999), and therefore, we decided to

include this as an individual level predictor in our analysis too. This relationship is in line

with so called, congruence hypothesis according to which evaluations of the domestic

political system serve as a basis for evaluation of supranational institutions like the EU. This

hypothesis has received support in several researches (de Vries and van Kersbergen, 2007;

Scheuer and van der Brug, 2007) and we assume that people who are satisfied with national

democracy will be more likely to show higher trust in European Commission.

The final individual level predictor included in the analysis refers to interest in politics. We

used a composite measure of interest in politics made out from Q1 (all items), Q2_1 and

Q2_3 (all items) as suggested in the scale and index construction recommendations. Higher

MYPLACE August 2014


interest in politics has positively correlated with support for European institutions (Arnold et

al., 2012), but this should not be equated with education because education has the

opposite effect. Education was not included in analysis because respondents in our sample

were still in education. All individual level variables were grand-mean centred.

Contextual variables

We included four contextual variables in the analysis. First refers to welfare state types

proposed by Kaariainen and Lehtonen (2006). An additional two contextual variables were

derived from The Quality of Government Institute database (Jan et al., 2013), one

representing a measure of society’s political functioning (government effectiveness), and

the other representing estimates of economic functioning (GDP per capita). Finally, we

included a widely used Corruption perception index. A brief description of these four

contextual variables is, presented in the appendix.

Results

The inclusion of individual level variables resulted in 18.7% decrease of the residual,

meaning that 18.7% of variance in trust towards the European Commission within certain

country can be attributed to a set of individual level predictors included. Chi-square test of

deviances (loglikelihood = 18919.746; df = 8) confirms that this is a highly significant

difference. Inclusion of individual level variables led to 46.2% of reduction in intercept

variance, meaning that almost half of the variance in trust in the European Commission

between countries can be attributed to the set of individual level predictors.

In the next step we included contextual variables, which led to 78.8% reduction of intercept

variance compared to the model containing only individual level predictors. This difference

in the fit of the model in comparison to a model containing only individual level predictors is

significant at .01 level as indicated by chi-square test of deviance (loglikelihood = 22.208;

df = 7).

Table 5.3.3 Variance components of regression models explaining trust in European

Commission

Empty model

(random intercept

only) – Model 0

With individual

level explanatory

var’s - Model 1

With country level

explanatory var’s -

Model 2

Trust in

European

commission


σ (country level) 0.464 0.249 0.053

-2 log likelihood 68919.482 49999.736 49977.529

df 3 11 18

MYPLACE August 2014


Out of three socio-demographic predictors included in the Model 1 gender and class were

significant at .001 and .01 level, respectively. Contrary to results demonstrated by Nelsen

and Guth (2000), women in our sample seem to have more trust in European Commission

than, male respondents. Young people who belong to lowest social class also indicated

lower trust in European Commission, as was expected. Considering individual level

indicators, self-positioning on the left-right scale, level of life satisfaction and satisfaction

with democracy were all positive predictors of trust in European Commission (p<.001),

indicating that a one unit increase on the 11 point left-right scale, life satisfaction scale, and

satisfaction with democracy scale is associated with 0.04, 0.12 and 0.28 increase in trust in

European Commission, respectively. On the other hand, a one unit increase in the

composite interest in politics scale was associated with 0.29 decrease in trust in the

European Commission.

In the next step (Model 2) we included four contextual variables out of which only

corruption perception was significant at .05 level. Higher numbers on CPI indicate lower

corruption perception, and therefore we can conclude that countries with lower

perceptions of corruption have more trust in the European Commission. Level of trust in the

European Commission was higher in countries with higher GDP level (p<0.001), but due to a

large deviation of the GDP measure the b weight for this predictor is low.

In Model 1 and Model 2 we demonstrated that young people who self-positioned

themselves as more right- wing on a left-right scale had more trust in the European

Commission. In the literature there is no clear association between self-positioning on left-

right scale and trust in institutions such as the European Commission. However, some

authors claim that the connection between left-right and attitudes towards European

institutions depends on the national context, more specifically on the type of welfare state

(Marks and Hooghe, 2003). It is hypothesised that countries with strong welfare institutions,

such as Nordic countries, perceive European institutions as an economic burden for their

budgets and that those on the right wing are more positively disposed toward their

institutions. In liberal countries where welfare state is weaker, such as United Kingdom, it is

hypothesised that those who are more inclined towards left could have more trust in

European institutions. As a prerequisite for running Model 3 we confirmed that slope

variance for left-right-trust varies across countries (Wald Z=2.125, p<.05). Chi-square test of

deviance confirmed significant increase in fit of this model in comparison to Model 2

(loglikelihood = 31.78; df = 3). In order to test these, assumptions we included two

interaction terms in our model, interactions between self-positioning on left-right scale and

a dummy variable specifying Nordic (Left_right_gmc*nordic_dummy) or liberal

(Left_right_gmc*liberal_dummy) countries. If there were a significant interaction term, we

would conclude that the slope of left-right ideology and trust varies across countries. As is

evident from the two last rows in Table 5.3.4 these interaction terms were not significant.

MYPLACE August 2014


Table 5.3.4 Determinants of trust in the European Commission Model 1 Model 2 Model 3

b p SE b p SE b p SE


Gender (1=M, 2=F) .140 *** .039 .141 *** .039 .141 *** .039

Class (0=lowest, 1=higher and

high)

.127 ** .040 .126 ** .040 .122 ** .040

Age -.004 .007 -.004 .007 -.005 .007

Left-right .040 *** .009 .040 *** .009 .036 .020

Social trust -.012 .009 -.013 .009 -.014 .009

Life satisfaction .119 *** .010 .119 *** .010 .122 *** .010

Satisfaction with democracy .281 *** .009 .281 *** .009 .279 *** .009

Interest in politics -.292 *** .023 -.293 *** .023 -.292 *** .023


Welfare state - Post-socialist -.259 .535 -.140 .555

Welfare state - Nordic -.330 .372 -.339 .385

Welfare state - Conservative -.138 .298 -.165 .308

Welfare state - Mediterranean .475 .404 .580 .419

Welfare state - Liberal ref. ref. ref. ref.

Government effectiveness -.586 .553 -.648 .574

Corruption perception index

(CPI)

.056 * .019 .060 ** .020

GDP per capita .000 ** .000 .000 * .000

Left_right_gmc *

nordic_dummy

.003 .054

Left_right_gmc *

liberal_dummy

.047 .073

***p<0.001, **p<0.01, *p<0.05

MYPLACE August 2014


5.4 Summary

In this chapter, we reported the findings of a cross-country analysis of youth attitudes

towards politicians and politics as well as trust ininstitutions.












scale.

The construction and testing of composite variables based on cross-country merged data

sets revealed that three items related to youth attitudes towards politics and politicians

constitute a one-dimensional scale of acceptable reliability. However, in order to reduce the

share of missing values we used only two items (Q6_2 and Q6_3) for the construction of the

‘Attitude towards politicians and politics’ scale, which also has higher reliability than the

three item scale.

Descriptive analyses revealed that young people in general have very negative attitudes

towards politicians and politics (mean=2.5, on scale from 0 to 8), whereas means scores by

locations range from 1.1 (in New Philadelphia Greece) to 4.6 in Odense Center (Denmark).

Thus, youth in almost all locations on average has negative attitudes towards politics and

politicians (with the exception of locations in Denmark and Finland where the average

attitude is neutral). The most negative attitudes towards politicians and politics were

characteristic for youth in Croatia, Slovakia, Portugal, Spain and Greece. Descriptive analyses

did not reveal significant differences in youth attitudes towards politics by locations within

countries.

With respect to analyses related to the construction and testing of composite variables

based on cross-country merged data about youth trust in 13 different institutions and

organisations we calculated two composite scores: overall trust in institutions (9 items) and

trust in national political institutions (3 items), both of which represent good scales.

Besides these composite results we also used one item measures towards specific

MYPLACE August 2014


institutions: trust in national parliament and trust in European Commission depending on

the topic of further analysis.

Descriptive analyses revealed that young people are, characterised by a low level of trust in

institutions. The overall mean for trust in institutions and organisations is 43.6 (on a scale

from 0 to 90), while the overall mean for trust in national political institutions is 12.5 (on a

scale from 0 to 30). In both cases only in five out of 14 MYPLACE countries do young people

in both locations show trust levels above the midpoint of the scale (in Finland, Georgia,

Denmark, and Germany (eastern and western)). The lowest levels of overall trust towards

institutions as well trust in national political institutions were characteristic for youth in

Greece, Croatia, Spain and Hungary (and Portugal in the case of trust in national political

institutions). The descriptive analyses did not reveal significant differences in young people’s

general level of trust in institutions or their level of trust in national political institutions by

locations within countries.

Regarding the trust in national parliament the results were similar, youth expressed on

average medium levels of trust in national parliaments (mean=4.5 on a scale from 0 to 10),

which varies by location ranging from 2.6 in Pescenica (Croatia) to 6.5 in Odense Center

(Denmark). At the country level, only five of the MYPLACE countries show a level of trust in

the national parliament above the midpoint of the scale (Denmark, Finland, Georgia and

both eastern and western Germany), while the lowest level of trust in the national

parliament is characteristic for youth in both locations within Croatia, Greece and Hungary.

Average levels of trust in the European Commission is slightly higher than in national

parliament (5.2 on a scale from 0 to 10), although still moderate, with the highest mean

score in Telavi ( Georgia, mean=6.6) and the lowest mean score at two locations in Greece

(Argyroupouli, M=3.5 and New Philadelphia, mean=3.7). However, more than half of the

MYPLACE countries showed levels of trust in the European Commission above the midpoint

of the scale (Denmark, Finland, Georgia, eastern and western Germany, Spain and Portugal).

The lowest level of trust in the European Commission is characteristic for youth in locations

within Greece, Croatia, Slovakia and Hungary. Significant differences between locations

within countries in young people’s level of trust in national parliaments, as well as in the

European Commission, were more an exception than the rule. In the case of national

parliaments, differences were evident in three countries (Latvia, Russia and UK) while in the

case of the European Commission differences were found in four countries (UK, Latvia,

Estonia and Finland).

In order to explore the multivariate determinants of trust in national parliaments as well

trust in the European Commission, taking into account the nested structure of our data set,

we employed multilevel modelling. Initial results confirmed the justification of multilevel

analyses since it was demonstrated that 17% of the variance of trust in national parliaments,

MYPLACE August 2014


and 7% of the variance of trust in the European Commission can be explained at the level of

countries. The percentage of variance which could be explained at the level of locations

within countries was much lower: 2% in case of trust in national parliaments; and 2% of the

variance in the case of trust in the European Commission. Such findings are in accordance

with the above and suggest relatively rare significant differences in youth levels of trust in

institutions between the two contrasting locations within countries. Thus, regarding the first

hypothesis according to which youth trust in national parliaments and trust in the European

Commission are differentiated by individual level characteristics as well as location and

country level characteristics, the above findings demonstrate that country level

characteristics are more important determinants of youth trust in national parliaments and

the European Commission than location level characteristics. Additionally they indicate

country level characteristics are more important determinants of trust in national

parliaments than trust in the European Commission. In further analyses we tested several

individual and country level characteristics as potential determinants of trust in national

parliaments as well as trust in the European Commission. Both regression analysis at the

individual level included the main socio-demographic variables (gender, age and social

class) and additional individual level variables, which represent more general dispositions

relevant in the context of trust in institutions such as social trust and life satisfaction, as well

as more politically relevant attitudes such as interest in politics (composite measure) and

satisfaction with democracy. Additionally, analyses of trust towards national parliaments

included a variable on ‘lack of care’ operationalised as agreement with the statement

Politicians are interested in young people like me. These individual level variables explained

29% of variance in trust in national parliaments within countries, and 19% of variance in

trust in the European Commission. This difference probably could be attributed to the fact

that the model with trust in the national parliament contained one more individual level

variable than the model with trust in the European Commission as a dependent variable,

which is, moreover, confirmed as the most important predictor of trust in the national

parliament (the variable concerning politicians’ ‘care’ about ‘young people like me’).

As the individual level determinants of trust in the national parliament analysis revealed, all

socio-demographic variables were significant8; higher trust in national parliament is more

characteristic for females as well as youth with higher and highest social class and younger

youth. In the case of trust in the European Commission the results were similar with regard

to gender and social class; however, age was not confirmed as an important independent

predictor of trust in the European Commission. Among other individual level variables, life

satisfaction, satisfaction with democracy and interest in politics are confirmed as significant

independent predictors of trust in national parliaments as well as trust in the European

Commission. This indicates that young people who are more satisfied with life in general,

and with democracy in their country, are particularly characterised by higher levels of trust

8 Due to the large N we are taking into account only variables which are significant on p<0.01

MYPLACE August 2014


in national parliaments as well as in the European Commission. However, it seems that the

role of interest in politics is different for the case of trust in national parliament and the

European Commission. Namely, in the case of trust in national parliaments higher levels of

trust were more characteristic for those who have greater interest in politics, while, in the

case of trust in the European Commission. those who have greater interest in politics

showed lower levels of trust. These findings might be verified by employing other

operationalisations of interest in politics such as one based not only on self-reported

political interest but also a measure of following politics in the media. Additionally, these

findings could be a result of the presence of an additional individual level variable in the

individual level model of trust in national parliaments (whether politicians are interested in

‘young people like me’).

In the second model we tested the effect of country level variables: welfare state,

corruption perception index and rule of law estimate in the case of trust towards national

parliaments; and welfare state, corruption perception index, government effectiveness and

GDP per capita in analyses of trust in the European Commission. We did not confirm

significant (p<.01) effects for these variables in the case of trust in national parliaments. In

case of the European Commission, only the effect of GDP was significant (p<.01), but this

was negligible.

Finally, in the third model we tested the interaction between one individual level variable

and one country level variable in order to test the hypothesis that the relationship between

individual level determinants of trust in national parliaments and the European Commission

depends on context (specific location or country). In the case of trust in national parliament

our hypothesis was that the effect of an individual’s perception of politicians’ interest in

young people (care about young people like me) as an individual level determinant of trust

in national parliament could depend on a country’s level of corruption (CPI). In the case of

trust in the European Commission our hypothesis was that the effect of left-right political

ideology as an individual level variable could depend on the type of welfare state. However,

we did not confirm these expectations. Thus, it might be concluded, that we did not confirm

the general hypothesis according to which the relationship between individual level

determinants of trust towards national parliament and the European Commission depends

on context (specific location or country). However, it should be stressed that this does not

mean that this general hypothesis is not correct; it only means that it is not confirmed by

those combinations of variables included in the models tested here.

MYPLACE August 2014


References

Armingeon, K., and Ceka, B. (2014). The loss of trust in the European Union during the great

recession since 2007: The role of heuristics from the national political system. European

Union Politics, 15: 82-107.

Arnold, C., Sapir, E.V., Zapryanova, G. (2012): ‘Trust in the institutions of the European

Union: A cross - country examination’ In: Beaudonnet, Laurie and Danilo Di Mauro (eds)

‘Beyond Euro - skepticism: Understanding attitudes towards the EU ’, European integration

online Papers (EIoP),Special Mini-Issue 2,Vol. 16, Article 8,

http://eiop.or.at/eiop/texte/2012-008a.htm.

Banaji, S., Buckingham, D., 2010. Young People, the Internet, and Civic Participation: An

Overview of Key Findings from the CivicWeb Project. International Journal of Learning and

Media 2, 15-24

de Vries, C.E. and van Kersbergen, K. (2007). Interests, Identity and Political Allegiance in

the European Union. Acta Politica, 42(2-3), 307-328.

Fisher, J., Van Heerde, J. and Tucker, A. (2010), Does One Trust Judgement Fit All? Linking

Theory and Empirics. The British Journal of Politics and International Relations, 12: 161–188.

doi: 10.1111/j.1467-856X.2009.00401.x

Hakhverdian, A., and Mayne, Q. (2012). Institutional trust, education, and corruption: a

micro-macro interactive approach. The Journal of Politics, 74(03), 739-750.

Henn, M., and Foard, N. (2012). Young people, political participation and trust in Britain.

Parliamentary affairs, 65(1), 47-67

Hooghe, M., and Wilkenfeld, B. (2008). The stability of political attitudes and behaviors

across adolescence and early adulthood: A comparison of survey data on adolescents and

young adults in eight countries. Journal of Youth and Adolescence, 37(2), 155-167.

Jan, T., Charron, N., Dahlberg, S., Holmberg, S., Rothstein, B., Sundin, P. and Svensson, R.

(2013). The Quality of Government Basic Dataset made from The Quality of Government

Dataset, version 15May13. University of Gothenburg: The Quality of Government Institute,

http://www.qog.pol.gu.se.

http://eiop.or.at/eiop/texte/2012-008a.htm

http://www.qog.pol.gu.se/

MYPLACE August 2014


King, D. C. (1997). The Polarization of American Parties and Mistrust of Government. In: J.S.

Nye, P. D. Zelikow and D. C. King (Eds). Why People Don’t Trust Government. Cambridge,

Mass.: Harvard University Press.

Marien, S. (2011). Measuring Political Trust Across Time and Space. In: Hooghe M., Zmerli S.

(Eds.), Political Trust. Why Context Matters, Chapt. 2. Colchester: ECPR Press, 13-46.

Marks, G. and Hooghe, L. (2003). National Identity and Support for European Integration.

Wissenschaftszentrum Berlin für Sozialforschung (WZB), 1-40.

McLaren, L. (2007). Explaining Mass-Level Euroscepticism: Identity, Interests and

Institutional Distrust. Acta Politica, 42(2-3), 233-251.

Mishler, W., and Rose, R. (2001). What are the origins of political trust? Testing institutional

and cultural theories in post-communist societies. Comparative political studies, 34(1), 30-

62.

Mondak, J.J. and Halperin, K.D. (2008). A Framework for the Study of Personality and

Political Behavior. British Journal of Political Science, 38(2), 335-362.

Muñoz, J., Torcal, M., and Bonet, E. (2011). Institutional trust and multilevel government in

the European Union: Congruence or compensation?. European Union Politics, 12(4), 551-

574.

Newton, K. (2013). Social and Political Trust. Norwegian Social Science Data Services

(http://essedunet.nsd.uib.no/cms/topics/2/3/)

Newton, K., and Norris, P. (2000). Confidence in Public Institutions: Faith, Culture, or

Performance. In S. J. Pharr and R. D. Putnam (Eds.), Disaffected Democracies: What's

Troubling the Trilateral Countries? (pp. 52-73). Princeton: Princeton University Press.

Norris, P. (1999). Institutional Explanations for Political Support. In: P. Norris, (Ed.), Critical

Citizens: Global Support for Democratic Governance. Oxford: Oxford University Press, pp.

217-236.

Roth F., Nowak-Lehmann D. F. and Otter, T. (2013). Crisis and Trust in National and

European Union Institutions — Panel Evidence for the EU, 1999 to 2012. European University

Institute.

Scheuer, A. and van der Brug, W. (2007). Locating Support for European Integration. In: W.

van der Brug and C. van der Eijk, (Eds). European Elections and Domestic Politics. Lessons

from the Past and Scenarios for the Future. Notre Dame. University of Notre Dame Press, 94-

116.

http://essedunet.nsd.uib.no/cms/topics/2/3/

MYPLACE August 2014


Schoon, I., and Cheng, H. (2011). Determinants of Political Trust: A Lifetime Learning Model.

Developmental Psychology, 47 (3), 619 631. doi: 10.1037/a0021817

Torney-Purta, J., Barber, C. H., and Richardson, W. K. (2004). Trust in Government-related

Institutions and Political Engagement among Adolescents in Six Countries. Acta Politica,

39(4), 380-406.

van der Meer, T. (2010). In what we trust? A multi-level study into trust in parliament as an

evaluation of state characteristics. International review of administrative sciences, 76(3),

517-536.

van der Meer, T. and Dekker, P. (2011). Trustworthy state, trusting citizens? A multilevel

study into objective and subjective determinants of political trust. In: S. Zmerli, and M.

Hooghe (Eds.), Political Trust: Why Context Matters (p. 95-116), ECPR Press.

White, C., Bruce, S., and Ritchie, J. (2000). Young people's politics: Political interest and

engagement amongst 14-24 year olds. Layerthorpe: York Publishing Services.

MYPLACE August 2014


Chapter 5: Appendix 1 Country-level variables

Government Effectiveness - Estimate (gov_eff) - Government Effectiveness combines

into a single grouping responses on the quality of public service provision, the quality

of the bureaucracy, the competence of civil servants, the independence of the civil service

from political pressures, and the credibility of the government’s commitment to policies.

The main, focus of this index is on “inputs” required for the government to be able to

produce and implement good policies and deliver public goods. Source: Quality of

Government Institute database (Jan et al., 2013)

Rule of Law – Estimate (rule_law) - Rule of Law includes several indicators which, measure

the extent to which agents have confidence in and abide, by the rules of society. These

include perceptions of the incidence of crime, the effectiveness and predictability of

the judiciary, and the enforceability of contracts. Together, these indicators measure

the success of a society in developing an environment in which fair and predictable

rules form the basis for economic and social interactions and the extent to which property

rights are protected. Source: Quality of Government Institute database (Jan et al., 2013)

GDP per Capita, PPP, Constant International USD (GDP_per_capita) - GDP per capita based

on purchasing power parity (PPP). PPP GDP is gross domestic product converted to

international dollars using purchasing power parity rates. An international dollar has the

same purchasing power over GDP as the U.S. dollar has in the United States. GDP at

purchaser's prices is the sum of gross value added by all resident producers in the

economy plus any product taxes and minus any subsidies not included in the value of

the products. It is calculated without making deductions for depreciation of fabricated

assets or for depletion and degradation of natural resources. Data are in constant 2005

international dollars. Source: Quality of Government Institute database (Jan et al., 2013)

Corruption perception index (CPI) - In addition to these three contextual variables taken

from Quality of Government Institute database (Jan et al., 2013), we have additionally

included a widely used Transparency International’s measure of corruption – Corruption

perception index (CPI). CPI scores and ranks countries based on how corrupt a country’s

public sector is perceived to be. Varieties of institutions participate in collection of data for

creation of this composite index. Since there is no meaningful way to assess levels of

corruption in different countries because they are usually comprised of illegal activities that

are, deliberately hidden, CPI is capturing perceptions of corruption of those in a position to

offer assessments of public sector corruption.

MYPLACE August 2014


Chapter 6. Political Activism

In this chapter we examine a set of questions related to youth political activism and civic

engagement. Voting behaviour is one of the most important indicators of the relationship

between citizens and political organisations, and is essential to the legitimisation of political

power (Viegas and Faria, 2003) as well as being fundamental in the support of the

democratic system. While in recent decades there has been a significant decrease in the

rates of electoral turnout in most European countries (http://www.parties-and-

elections.eu/), of most significance to MYPLACE is the fact that throughout Europe it is

among young people that participation tends to be weakest (Fieldhouse et al.,2007;

Goerras, 2007), falling short of the level of participation among older age cohorts. Research

on this issue has shown that young people who do not exercise their right to vote the first

two opportunities for which they are eligible will probably turn into recurrent abstainers

(EACEA, 2013). This data supports the urgency of bringing this issue into the centre of

debate in order to identify the causes and to identify adequate solutions to this problem. In

this context, the objectives and methodologies adopted by MYPLACE can make an

important contribution to this discussion, encompassing 14 countries with very different

cultural, political and socio-economic backgrounds. Framing electoral behaviour with other

political and civic participation practices thus gives MYPLACE a vantage point onto a wider

understanding of this phenomenon.

In the last few decades, non-electoral participation activities have been increasingly used to

explain citizens’ political activism and particularly, the political activism of young people.

Several studies have demonstrated that electoral participation has decreased in younger

generations (Franklin et al., 2004; Putnam 2000; Pirie and Worcester 1998). However, and in

parallel, some other forms of participation have taken a growing role in explaining how

young people relate to politics. In particular, in most European countries an increase in

protest actions, political consumerism or Internet political activities have been detected

(Norris, 2004; Zukin et al., 2006; Dalton 2008; Michelletti, Follesdal and Stolle 2003; Stolle

and Hooghe, 2005). Young people seem to be diversifying the mechanisms that they use to

participate as they have a wider conceptualisation of the political (Henn et al., 2002; O'Toole

2003; Manning 2010). Consequently, in order to understand how young people behave

politically, it is necessary to have a complete battery of different political actions. One of

the analytical departing points of the MYPLACE project has been to approach the political

views and involvement of young people in a comprehensive way. Therefore, it was

necessary to consider a large number of political activities from different natures in order to

allow different approaches to political activism to emerge. To do so, the MYPLACE survey

includes a complete battery of 20 different non-electoral political actions (Q16) where

different dimensions of political activism can be found, such as classic institutional

participation actions (like volunteering in an election campaign or contacting politicians),

http://www.parties-and-elections.eu/

http://www.parties-and-elections.eu/

MYPLACE August 2014


confrontational protest (like participating in a violent political event, occupying buildings or

blocking street), Internet activities (like uploading political material to the Internet, writing

in a blog) or expressive activities (like wearing a badge with a political message), among

many others.

In previous analyses from the MYPLACE project, this battery of political actions has already

been analysed for each country. These national reports (deliverables D4.5) have

demonstrated the great diversity of actions that young people use to participate. However,

these reports have also pointed out a great deal of diversity on both the level and the type

of participation among the European localities included in the MYPLACE project. In this

chapter, we want to go delve deeper into these differences.

6.1 Voting behaviour (Q8-Q13)

MYPLACE findings show that less than half (43.1%) of a total of 16,727 respondents

reported casting a vote in their country’s last national elections. Moreover, only 31% in a

total of 15,469 have voted in the last local elections. From those who did not vote in

national elections, 67.6% out of a total of 9,183 respondents were not eligible, 12.8% stated

that they would have liked to have voted but were unable to do so on that day, 6.4%

decided not to vote because there was no party aligned with their views, for 8.7% voting or

not voting was seen as equally pointless and finally 4.5% did not vote to show their

dissatisfaction with politicians and parties. Regarding local elections, the results are similar:

65% in a total of 10,774 respondents were not eligible to vote, 16% would have liked to

have voted but were unable to do it on the day, 6.7% decided not to vote because there was

no party aligned with their views, for 8.9% voting or not voting is equally pointless and 3.4%

did not vote to show their dissatisfaction with politicians and parties.

These results show that the main reason for not casting a vote in the last elections (national

or local) was simply not being eligible. This means that to capture meaningful motives for

not voting among the ones that had the right to do so requires the exclusion of the non-

eligible from the analysis. After this, it is admissible to say that 69% of the respondents

declared they went to the polls to vote in the last national elections. However, turnout rates

differ significantly from country to country. Denmark (93%), eastern Germany (84%) and

Spain (80%) have the highest turnout rates while Portugal (59%), Russia (50%) and UK (44%)

have the lowest. In local elections turn outs are even lower: 54% of eligible respondents

report that they voted in the last local elections. Spain (76%), Greece (70%) and eastern

Germany (69%) have the top positions; Russia (42%), Slovakia (36%) and UK (33%) were

bottom.

MYPLACE August 2014


According to the literature, there are several variables that have to be considered in the

analysis of voting behaviour. The first set encompasses socio-demographic factors,

portraying the first picture of our research target and possible predictors of our dependent

variable – vote behaviour. Age is one of most meaningful variables in respect to voting

behaviour or elections turnout. Research has indicated a positive relationship between age

and turnout (Henn et al., 2002; Magalhães, 2001; Freire, 2001; Fieldhouse et al., 2007). In

other words, it is among the youngest that it is possible to find higher rates of electoral

turnout. On the other hand, gender displays a weaker or absent relation with turnout rates

(Viegas and Faria, 2004). However, while not statistically significant, women show higher

turnout rates (Fieldhouse et al., 2007; Viegas and Faria, 2004), and it is always important to

understand the behaviour of this independent variable.

Another important independent variable is education. The literature demonstrates that less

educated people tend to vote less frequently (Viegas and Faria, 2004; Fieldhouse et al.,

2007). Given that a large proportion of our respondents are, still in education we might

predict a negative relationship between education and voting behaviour among our sample.

Being religious is one of variables that is also positively related with voting behaviour,

according to Fieldhouse et al.(2007) respondents that declare themselves religious are more

likely to vote. This tends to happen because religious practices are considered a form of

social participation; therefore those who have a higher level of social participation in society

tend also to have higher political participation (Jonas-Correa and Leal, 2001).

Regarding employment status, Fieldhouse et al.(2007) reveal that there is no impact of

employment status on turnout rates; however, Viegas and Faria (2004) found contradictory

results, with active (employed) people tending to abstain more often. Considering that

youth unemployment is an important issue to Europe and to a significant proportion of

MYPLACE countries, it is essential to analyse how this independent variable may influence

turnout rates. Also important is social class. According to Flanagan and Levine (2010),

parents of high socio-economic status pass on to their children political awareness and

access to community and education resources, which leads to increased political

participation. This may have an important effect among MYPLACE respondents. Regarding

ethnic minorities, it is well documented that minorities tend to have less political

participation and lower electoral turnout. Therefore we expect to find similar results among

MYPLACE respondents.

Despite the significance of socio-economic variables to the understanding of the levels of

political participation, several other variables should also be considered. However, the

choice of these variables depends on the theoretical model adopted to explain the general

political participation and voting behaviour in particular. Among the various theories that

have emerged to explain this phenomenon, three have been highlighted in the literature

MYPLACE August 2014


(Fieldhouse et al., 2007): rational choice theory (or general incentives), social capital theory

and the theory of civic voluntarism.

The theory of rational choice, accentuates the utility character of the vote. The vote is seen

as an investment in which participants expect to get some benefit from the victory of the

candidate they support. Social capital is based on the assumption that volunteer activities

encourage trust in others and encourage political activism (Putnam, 1993). Studies based on

the importance of social capital to voting behaviour try to observe the relatively minor

involvement of young people in civil society and also a lesser connection to the ideals of

democracy and democratic institutions, compared with older age groups (Huggins, 2001).

Young people are also the less trusting in governments and political institutions (Nye, 1997).

The last model - civic voluntarism - is based on the assumption that citizens with more

resources participate most in politics, i.e., young people’s low participation in political life is

associated with their lack of resources to establish connections with political institutions

and the sense of duty of the electorate as a whole. There are a number of studies that

suggest that young people in many countries are decoupled from the central aspects of

politics (Wattenbergm, 2002; Park, 2000). The model of civic voluntarism groups a set of

concepts in particular the level of involvement of young people in society, the political

efficacy (in other words, the perception that common citizens can change politics); the

concept of civic duty and finally the level of involvement of the respondents in politics are

the key variables to support this theoretical model. In order to analyse voting behaviour we

have decided to choose this last model; however, it was necessary to do make some

adaptations, as some variables such as civic duty were not included in the questionnaire.

As important as studying the individual characteristics of voters is the context in which they

are immersed. MYPLACE respondents are, nested in localities, which in turn are located in

countries. In this sense, it is important to consider other exogenous variables that may

influence voting behaviour, such as welfare state regimes and differences among political

systems. In order to analyse the impact of these factors on voting behaviour we have

introduced some contextual variables such as welfare regime, number of representation

cameras and type of parliamentary representation. According to Lister (2007) the model of

welfare state influences social norms and hence individual behaviour. In more egalitarian

societies with a Universalist Welfare State, it is predictable to find higher levels of political

participation. Some studies point out that countries with bicameral systems tend to have

higher rates of turnout due to the competition between the two bodies. Regarding types of

representation, the literature states that countries with proportional representation reveal

higher rates of turnout. This may happen because voters can be sure that they will be

represented in parliament even if it is only the opposition bench.

Analysing such variables entails considering a new level of analysis, which calls for a

multilevel modeling, balancing contextual variables with individual voter characteristics. In

MYPLACE August 2014


this case, we will consider two levels; the individual level (Level 1) and the localities level

(level 2) once the 16,935 individuals are nested within 30 localities. In this report, we are

considering two types of elections - national and local. In this sense, we opted to use,

whenever possible, the same predictor variables in order to evaluate if these variables have

the same impact on different types of elections. The only difference being the contextual

variables as we use characterise the political system according to the type of elections we

were analysing - national and local.

MYPLACE August 2014


Q8: Voting behaviour in the last national election (including those not eligible)


Nominal variable with two categories (Yes/No).The locations are

ordered by percentage of turnout in last national elections.

Graph 6.1a: Voting behaviour in last national election by location

Narrative

The overall percentage of turnout in the last national elections was

43% among a total of 16,727 respondents. This varies by location,

with the highest score Odense Center (Denmark) with 77.1%, and the

lowest score Nuneaton (UK) with 16.6%.


eastern Germany and Croatia all in the top third, with higher levels of

turnout. The UK, and Russia are in the bottom third with lowest

turnouts.

Countries have similar scores between the two locations; the

exceptions were Finland (Kuopio and Lieksa & Nurmes), with a

difference of 42.0% between the two locations, Estonia (Tartu &

Narva area) with 19.0%, Latvia (Agenskalns and Forstate & Jaunbuve)

with 15.3% and Portugal (Lumiar and Barreiro) 15.2% difference.

MYPLACE August 2014


Q10: Reasons for not voting in the last national election


The ‘Reasons to not vote in last national elections’ question consisted of a

nominal variable with five options of response:

I was not eligible

I would have liked to have voted but was unable to do so on the day

I decided not to vote because in this election there was no party

that aligned with my views

For me, voting or not voting is equally pointless

I did not vote to show my dissatisfaction with politicians and parties

The locations are ordered by the percentage of the most frequent category

(I was not eligible).

Graph 6.1b: Reasons for not voting inlast the national election by

location

Narrative

From 9,183 respondents that have not voted in the last national election,

the vast majority (67.6%) did not do so because they were ineligible.

Lieksa & Nurmes (Finland), Odense East (Denmark) and Vic (Spain) reveal

the higher proportion of respondents that were not eligible to vote at the

time of last national election (83%, 80% & 79% respectively).

Georgia is the country with lowest percentage of respondents that were not

eligible (Telaviv, 51% & Kutaisi, 43%). In this country, 45.5% said that they

would have liked to have voted but were unable to do so on the day.

83

28

69

20

72

81

77

61

92

51

64

51

56

47

83

36

84

74

31

65

62

55

52

40

52

56

49

44

41

31

3

3

5

3

4

6

8

8

4

7

9

4

8

8

15

7

12

15

4

15

4

13

13

10

13

22

24

8

32

36

5

1

5

1

6

3

8

4

10

4

4

6

4

5

3

4

5

3

6

6

12

8

5

5

10

8

5

13

4

2

6

2

5

1

4

12

3

6

7

3

4

6

4

2

15

3

18

15

4

9

12

10

7

10

9

16

15

9

2

1

4

1

4

0

6

2

4

1

8

1

4

3

7

3

4

2

5

2

2

4

7

4

9

3

4

3

1

12

2

1

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Lieksa & Nurmes (FIN)

Odense East (DEN)

Vic (SPA)

Odense Center (DEN)

New Philadelphia (GRE)

Forstate & Jaunbuve (LAV)

Bremen (GER-W)

Agenskalns (LAV)

Barreiro (POR)

Jena (GER-E)

Sant Cugat (SPA)

Podsljeme (CRO)

Argyroupouli (GRE)

Rostock (GER-E)

Narva area (EST)

Bremerhaven (GER-W)

Nuneaton (UK)

Coventry (UK)

Kuopio (FIN)

Lumiar (POR)

Sopron (HUN)

Tartu (EST)

Rimavska Sobota (SLO)

Pescenica (CRO)

Ozd (HUN)

Kupchino (RUS)

Vyborg (RUS)

Trnava (SLO)

Telavi (GEO)

Kutaisi (GEO)

I was not eligible

I would have liked to have voted but was unable to on the day

I decided not to vote because in this election there was no party that aligned with my views



MYPLACE August 2014


Q11: Voting behaviour in the last local election (including those not eligible)


Nominal variable with two categories of response (Yes/No). The

locations are ordered by percentage of turnout in last local elections.

Graph 6.1c: Voting behaviour in the local election by location

Narrative

The overall percentage of turnout in the last local election was 31.1%

from a total of 16,469 respondents. This varies by location, with the

highest percentage at Rostock (eastern Germany) with 55.6% of

turnout and Odense Center (Demark) with 50.7%, and the lowest

percentage at Nuneaton (UK) with 14.1% and Forstate & Jaunbuve

(Latvia) with 15.0%.

Eastern Germany (52.2%), Denmark (46.9%) and western Germany

(41.9%) had the highest percentage of turnout in local elections.

On the other hand, UK (17.8%), Slovakia (18.6%) and Latvia (20.9%)

clustered in the bottom third with lower levels of turnout in local

elections.

Most countries had similar percentage between the two locations;

exceptions were; Finland (Lieksa & Nurmes and Kuopio) with 25.9%

difference between the two locations, Portugal (Lumiar and Barreiro)

11.4% and Greece (New Philadelphia and Argyroupouli) 10.6%.

56

51

49

45

43

43

42

41

40

35

32

32

32

32

31

29

29

28

27

26

26

26

25

22

19

18

18

16

15

14

44

49

51

55

57

57

58

59

60

65

68

68

68

68

69

71

71

72

73

74

74

74

75

78

81

82

82

84

85

86

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Rostock (GER-E)

Odense Center (DEN)

Jena (GER-E)

Bremen (GER-W)

Kuopio (FIN)

Odense East (DEN)

Vic (SPA)

Sant Cugat (SPA)

Argyroupouli (GRE)

Bremerhaven (GER-W)

Ozd (HUN)

Pescenica (CRO)

Narva area (EST)

Sopron (HUN)

Vyborg (RUS)


Tartu (EST)

Lumiar (POR)

Agenskalns (LAV)

Kutaisi (GEO)

Podsljeme (CRO)

Telavi (GEO)

Kupchino (RUS)

Coventry (UK)

Trnava (SLO)



Barreiro (POR)


Nuneaton (UK)

Yes No

MYPLACE August 2014


Q13: Reasons for not voting in the last local election


There were five ‘Reasons to not vote in last local elections’

represented in a nominal variable. The locations are ordered by the

percentage of the most frequent category (I was not eligible).

Graph 6.1d: Reasons for not voting in the local election by location

Narrative

From 10,774 respondents that did not vote in the last local elections

the majority (63.6%) did not vote because they were not eligible;

15.9% were unable to do it on the day; 6.7% decided not to vote

because there was no party that aligned with their views; 8.9%

believed that voting or not voting is equally pointless and 3.6% did not

do it because they were disappointed with politicians and parties.

Barreiro (Portugal), Lieksa & Nurmes (Finland) and Forstate &

Jaunbuve (Latvia) reveal the higher proportion of respondents that

were not eligible at the time of last local election (82%, 82% and 81%

respectively).

Georgia (Kutaisi and Telavi) had the highest percentage of

respondents that did not vote because they were unable to on the

day even though they were eligible to do so.

82

61

85

57

57

68

57

59

70

66

52

60

59

26

52

37

46

21

56

48

51

36

46

44

37

20

35

24

20

35

3

3

8

4

6

7

5

5

13

10

5

11

10

8

25

13

6

6

16

14

16

12

32

18

16

15

22

18

12

19

5

3

4

4

3

5

4

5

3

4

3

5

7

3

3

6

10

3

4

5

10

8

2

8

9

3

3

6

7

5

6

5

7

4

4

3

3

4

5

4

7

7

10

4

1

3

9

4

16

15

6

6

1

13

8

3

13

3

6

19

4

2

1

3

3

4

5

3

1

4

3

2

3

1

2

1

5

1

4

2

6

3

1

3

3

2

1

2

2

4

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Barreiro (POR)



Sant Cugat (SPA)

Vic (SPA)


Argyroupouli (GRE)

Podsljeme (CRO)

Agenskalns (LAV)


Pescenica (CRO)

Lumiar (POR)

Tartu (EST)

Odense East (DEN)

Telavi (GEO)

Jena (GER-E)

Sopron (HUN)

Odense Center (DEN)

Nuneaton (UK)

Coventry (UK)

Trnava (SLO)

Bremen (GER-W)

Kutaisi (GEO)

Kupchino (RUS)

Ozd (HUN)

Bremerhaven (GER-W)

Vyborg (RUS)

Rostock (GER-E)

Kuopio (FIN)

Narva area (EST)

I was not eligible

I would have liked to have voted but was unable to on the day

I decided not to vote because in this election there was no party that aligned with my views



MYPLACE August 2014


6.2 Closeness with political parties (Q14-Q15)

Parties are an important intermediary between citizens and the state. They have a

fundamental role in democracy as they allow citizens to socialise political issues, and they

aggregate preferences, organise ideas, practices and priorities around ideological

orientations (Hoogle and Kern, 2013). However, in recent years their role in the political

system has changed. But even more relevant than the change in the importance of parties in

the political system was what has changed in respect to the role of citizens in the

development of party activities. Parties are increasingly professionalized and less dependent

on citizens, financially and logistically. The emergence in Europe of several cases of

independent candidates who are not, affiliated with a particular party has gained relevance

in the political landscape. These changes that have occurred in recent decades raise

questions about the importance of the proximity between citizens and parties and the role

of parties in the democratic system.

Our findings show that young people are the group with the most tenuous connection to

political parties. The decrease of importance of formal linkages to political parties widens

the interest of studying the informal relations between citizens and parties and the degree

of closeness with political parties. Although closeness with political parties or party

identification is not a sine qua non condition to political and electoral participation, it may

consist on a crucial indicator of the predisposition to exert both (Blaís et al., 2001).

In this report, besides the importance of socio-demographic variables we intend to verify

the nature of the relationship between political participation, political efficacy and trust in

political systems influence, closeness to a party and in which way this influence take place.

We predict that political participation, political efficacy and trust in political systems has a

positive impact on the closeness to political party.

MYPLACE August 2014


Q14: Closeness with political parties


Nominal variable with two categories of response (Yes/No). The

locations are ordered by percentage of respondents that declare

feeling close to a particular political party.

Graph 6.2: Closeness with political parties by location

Narrative

From a total of 15,957 respondents, more than half of respondents

(59.1%) declared that they did not feel close to any political party.

Denmark (66.6%), Spain (60.4%) and western Germany (57.8%) are on

the top third of countries with the highest proportion of young people

that feel close to a particular political party.

At the opposite extreme, Hungary (18.6%), Croatia (26.5%) and Latvia

(31.0%) are the countries with lowest percentage of young people

that feel close to a particular political party.

Most countries had similar percentages between the two locations;

notable exceptions were; Portugal (Lumiar and Barreiro), with a

difference of 22.8% between the two locations, and Latvia

(Agenskalns and Forstate & Jaunbuve) with 16.8%.

68

65

64

60

56

56

55

53

49

45

43

43

42

39

39

39

39

39

39

38

37

36

36

32

27

26

22

21

20

17

32

35

36

40

44

44

45

47

51

55

57

57

58

61

61

61

61

61

61

62

63

64

64

68

73

74

78

79

80

83

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Odense East (DEN)

Odense Center (DEN)

Sant Cugat (SPA)

Bremen (GER-W)

Vic (SPA)

Jena (GER-E)

Bremerhaven (GER-W)

Rostock (GER-E)

Kuopio (FIN)

Telavi (GEO)

Lumiar (POR)

Tartu (EST)

Coventry (UK)

Agenskalns (LAV)

Argyroupouli (GRE)



Narva area (EST)


Trnava (SLO)

Nuneaton (UK)

Kutaisi (GEO)

Kupchino (RUS)

Vyborg (RUS)

Podsljeme (CRO)

Pescenica (CRO)


Ozd (HUN)

Barreiro (POR)

Sopron (HUN)

Yes No

MYPLACE August 2014


6.3 Non-electoral political activities (Q16)

To analyse the political behaviour of young people beyond electoral participation, the

MYPLACE survey includes a battery of 20 different political activities. Respondents were

asked if they had undertaken each of these activities once, twice or three or more times in

the last 12 months.

We first start analysing an indicator that includes the participation in all the 20 items. This

variable indicates the global level of non-electoral participation on a 0-60 scale. However a

factor analysis, presented below, suggests that we can identify different dimensions of

political activism that are useful for understanding differences among locations.

Table 6.3 Factor analysis on political activities

Component

1 2 3 4

Volunteered in an election campaign 0.005 0.754 0.090 -0.038

Contacted a politician (e-mail / phone / SMS / letter /

fax etc) 0.402 0.525 -0.027 0.086

Attended a public meeting dealing with political issues 0.536 0.411 0.197 0.029

Signed a petition 0.681 0.125 0.083 0.057

Collected signatures 0.153 0.556 0.185 0.037

Given a political speech 0.198 0.519 0.036 0.178

Distributed leaflets with a political content 0.114 0.736 0.203 0.072

Boycotted and bought certain products for political,

ethical or environmental reasons 0.704 0.055 0.117 0.150

Written political messages or graffiti on walls 0.165 0.031 0.024 0.760

Worn a badge with a political message 0.376 0.382 0.324 -0.032

Participated in a demonstration 0.353 0.145 0.713 0.000

Participated in a strike 0.084 0.041 0.756 0.072

Donated money to support a political group or

organisation 0.339 0.270 0.079 0.137

Written an article -in an organisation journal, a blog- 0.451 0.261 -0.001 0.237

Written or forwarded a letter/an email with a political

content 0.632 0.214 0.163 0.027

Participated in a violent political event 0.009 0.143 0.220 0.717

Occupied buildings or blocked streets/railways -0.013 0.133 0.592 0.328

Participated in a flash mob 0.201 0.179 0.402 0.094

Uploaded political material to the internet 0.526 0.134 0.317 -0.064

Voted in student union elections 0.350 0.104 0.303 -0.143

MYPLACE August 2014


From these results, we constructed three different variables for three dimensions of

participation (mixing component 3 and 4).

Private individual participation (component 1) - It groups the less intense forms of

participation that usually occur in a private sphere. To do these types of actions, the

individual is not exposed to the public, so they take place in a more anonymous

environment. It is a 0-18 scale with a Cronbach Alpha of 0.73

Public traditional participation (component 2)- It differs from the former type of

participation because these actions are more exposed to the public. They are also

traditional ways of participation that are usually oriented to representative

institutions. It is a 0-15 scale with a Cronbach Alpha of 0.69.

Protest participation (components 3 and 4)- We have joined together components 3

and 4 because they both refer to protest actions. Component 4 captures the more

extreme and confrontational protest actions. It is a 0-18 scale with a Cronbach Alpha

of 0.62.

MYPLACE August 2014


Q16: Non-electoral participation (Global participation indicator)


This variable is a 0 to 60 scale where 0 means not having done any of

the 20 political actions proposed in the questionnaire in the last 12

month. As a scale, it has a good reliability with a Cronbach Alpha of

0.84. It has to be said that the variable has an asymmetric positive

distribution where the mode is 0. That means that the majority of the

respondents have not done any of these political actions.

Graph 6.3a: Non electoral participation (Global participation

indicator)

Narrative

The results show a great difference between locations: Sopron and

Ozd in Hungary present a level of participation 25 times lower than

the most participative locations like Jena and Rostock in eastern

Germany and Sant Cugat in Spain.

As a general trend, it could be said that locations in western Europe

tend to present a higher level of participation than locations placed in

eastern post-communist countries. Also, as a general consideration, it

seems that locations in the same country have a similar level of

participation when comparing with locations in other countries.

However, there are some remarkable exceptions. In Finland, for

example, the participation in Kuopio, a university city, is double the

participation than in both Lieksa and Numes, which are in very rural

areas. There are also remarkable differences in Portugal, United

Kingdom, Russia, Estonia and Latvia.

MYPLACE August 2014


Q16: Private individual participation


This variable is on a 0 to 18 scale. The scale captures whether the respondent has or

has not undertaken once, twice or three or more times the following political

activities: Attended a public meeting dealing with political issues; signed a petition;

boycotted and/or bought certain products for political, ethical or environmental

reasons, written an article (for example, in an organisation journal or a blog),

written or forwarded a letter/email with political content; and uploaded political

material to the Internet. The Cronbach Alpha for this scale is 0.73 and it also has an

asymmetric positive distribution.

Graph 6.3b: Private individual participation

Narrative

This dimension of participation is clearly the most used by young people in all the

locations. If we look at the results for Graph 6.3a (the global participation indicator)

and we compare them with the results for private individual participation, we can

see that in all the locations in the analysis, this kind of participation represents more

than the half of the global participation in political actions. Therefore, these kinds of

political activities that are carried out in a more private environment and at the

individual level, are the more common. This is because these activities are also the

ones that require a lower level of involvement in terms of costs and also, in terms of

being socially exposed. The other dimensions of participation analysed after, have

significantly lower levels, as they require more time and involvement. It could be

said that this private individual participation represents the first step in getting

involved while public traditional participation and protest are used for young people

that are more deeply engaged. Regarding the differences between locations, they

are very similar to what we have seen for the global participation indicator:

locations in western Europe present a higher level of participation and, with some

exceptions, locations in the same countries do not differ a considerably.

MYPLACE August 2014


Q16: Public traditional participation


The public traditional participation variable is on a 0 to 15 scale. The scale captures

whether the respondents have not undertaken or have undertaken once, twice or

three or more times the following political activities: volunteered in an election

campaign; contacted a politician; collected signatures, given a political speech and;

distributed leaflets with a political content. The Cronbach Alpha for this scale is 0.69

and it also has an asymmetric positive distribution.

Graph 6.3c: Public traditional participation

Narrative

The public individual participation is the dimension of political action analysed with

a lower level of participation. From a 0-15 scale, in any location the average

participation arrives at 1. That means that the vast majority of young people in all

these locations have not done any of these activities in the last year. Thus, MYPLACE

data confirms the weak propensity of young people to participate through

traditional institutional mechanisms (Putnam 2000; Pirie and Worcester 1998).

However, it is interesting to highlight that this kind of participation is the one that

presents less variation across locations. It seems that this is a stable form of political

action and it does not depend on the contextual factors as other forms do.

If we compare the position of the different locations, it appears that the fact that

locations in western countries tend to be more participative than locations in post-

communist Europe, is weaker here than in other forms of participation; Vyborg and

Kupchino, in Russia, show a high level of public traditional participation compared to

other countries, while Barreiro, in Portugal, New Philadelphia, in Greece, or Lieksa

and Nurmes, in Finland, present a very low level of this dimension in political action.

MYPLACE August 2014


Q16: Protest action


The protest variable is on a 0-18 scale. The scale captures whether the

respondent has or has not undertaken once, twice or three or more times

the following political activities: participated in a demonstration;

participated in a strike; participated in a violent political event; occupied

buildings or blocked streets/railways, written political messages or graffiti

on walls and; participated in a flash mob. The Cronbach Alpha for this scale

is 0.62 and it also has an asymmetric positive distribution.

Graph 6.3d: Protest action

Narrative

If public traditional participation was the dimension of political action with a

lower variation across locations, protest seems to be where we find greater

differences across locations. In some cases, protest represents a significant

part of the global participation (as is the case for the Spanish and Danish

locations) and in some others it has been really difficult to find any young

people active in protest actions. Therefore, protest action seems to be more

related to contextual factors than to other forms of participation.

Locations in Mediterranean welfare states (particularly those in Spain and

Greece) show a high level of protest activity, and again, countries with a

lower level of this kind of participation are locations placed in eastern

European countries. Here again, locations in the same country present a

very similar level of protest with Finland being the only exception.

MYPLACE August 2014


6.4. Multi-level Modelling

6.4.1 Model 1: Voting behaviour – National elections

According to prior studies, socio-demographic variables have an impact on voting behaviour. It is

among peripheral groups that it is possible to observe the highest rates of abstention (Viegas and

Freire, 2007). In other words, it is among the least educated, unemployed, lower social classes and

ethnic minority groups that we may observe lower turnout rates. The literature also shows other

variables with significant impact on voting behaviour. We have selected two of them: political

efficacy and political participation. Political efficacy refers to the perception that ordinary people

can influence political decisions. Therefore, those who have higher levels of political efficacy tend

to have higher turnout rates. In this report, this variable results from an index obtained through

the average answer to a set of items under the question ‘How effectively, do you think, the

following actions can influence politics in your country?’

Political involvement also has a positive impact on vote behaviour. The literature says that it is

between groups with higher levels of political involvement that it is possible to find higher turnout

rates. Political involvement was obtained through an index of variables used in a MYPLACE

previous report – WP6: Mapping Activism.

Besides individual variables, other contextual variables were considered, such as welfare state

regime, different type of parliament representation and cameral organisation. According to the

literature, countries with proportional representation tend to register higher levels of turnout.

This is also true to bicameral regimes. The welfare state regimes also can impact turnout rates.

Since more egalitarian societies favour the voting behaviour.

To sum up, we have predicted that age will have a positive impact on turnout. We have also

predicted that it is between young people with low education levels belonging to lower social class

families and minority groups and unemployed that abstention rate is higher. One other

assumption is that political efficacy and political involvement are positive influences on turnout

rates. Regarding contextual variables, and in line with the literature, we predict that countries with

proportional representation and bicameral regimes will show higher levels of electoral turnout.

Regarding the impact of welfare state regimes our hypothesis is that liberal regimes will have a

negative influence on turnout rates, once this regime emphasises the responsibility of individuals

for themselves and a weaker role of government and political decisions on people’s lives.

As said before, MYPLACE sample is hierarchically structured, since the respondents are nested into

localities and localities are nested in countries. This structure requires a more complex analysis of

MYPLACE August 2014


the data. In this way, it was chosen to pursue a multilevel approach. Results from multilevel

regressions suggest that the intercept variance varies between localities (σ). According to the

empty results (Z=3.588, p<.001) we can assume statistically significant variability in intercepts

across localities, which justifies developing a multilevel model. The Inter Class Correlation (ICC)

describes the proportion of variance that lies between localities relative to total variance. In this

case, the results suggest that 13.5% of the variability in voting behaviour lies between localities

(ICC= 0.135). When we introduce individual level explanatory variables we can see a significant

reduction of interclass class correlation (ICC=0.103), indicating that 10.3% of variance between

localities is still unexplained. The Z-test shows that the variance is statistically significant (z=3.513,

p<.001) suggesting that we could develop a model with contextual variables – a third model. In

fact the introduction of country level variables allowed a significant reduction on the unexplained

variance between localities to 7.2% (ICC=0.072).

Table 6.4.1.1 Variance components of regression models explaining vote behaviour

Empty model

(random

intercept

only)

With

individual

level

explanatory

var’s

With country

level

explanatory

var’s

Vote in national

elections σ (Location level) 0.513 0.377 0.254

Z

3.588

(.000***)

3.513

(.000***) 3.016 (.003**)

AIC 39495.542 40907.833 40946.325

BIC 39502.613 40914.902 40953.394

-2 log pseudo

loglikelihood 39493.541 40905.833 40944.325

***p<0.001, **p<0.01, *p<0.05

The regression coefficients of the last model allow some conclusions on the contribution of each

predictor variable to voting behaviour. According to the initial hypothesis, age increases log odds

of voting, and this positive relationship between age and voting behaviour corroborates the

findings obtained through other research on the subject. Regarding gender, the initial hypothesis

was not confirmed. The findings demonstrate a relationship between gender and voting

behaviour. Opposite to what has been observed in previous studies, among our sample being male

increases by 0.064 the log odds of voting in national elections.

In respect to education, we have observed that ISCED 0-2 education attainment reduces the log

odds in 0.091 on voting in the national elections, compared with respondents with higher levels of

education, namely ISCED 7-8. Despite the lower log odds of other levels of education when

compared to the higher level, this reduction is not so pronounced as the one observed in ISCED 0-

2 level. The results reveal that increasing levels of education will increase the turnout rates.

MYPLACE August 2014


The results also show that being employed or still in education increases the log odds of voting

when compared with unemployed groups. Being in education increases 0.479 times the log odds

of voting in national elections and an employed respondent has 0.420 times more log odds of

voting than an unemployed respondent.

Table 6.4.1.2. Parameter estimates of regression models explaining vote behaviour in national

elections (n2=30, n1=8703)

Empty model

(random

intercept only)

With individual

level explanatory

var’s

With country level

explanatory var’s

SE SE SE

Individual-level predictors

Intercept 0.911*** .000 -1.707** .002 -2.813*** .000

Age 0.065*** .000 0.064*** .000

Gender (male) 0.202** .002 0.201** .002

Education: ISCED 0-2 -0.089*** .000 -0.901*** .000

Education: ISCED 3-4 -0.270 .249 -0.276 .233

Education: ISCED 5-6 -0.109 .596 -0.110 .588

Education: ISCED 7-8 (ref. cat.)

Religion 0.041*** .000 0.041*** .000

Activity status: Employment 0.418*** .000 0.420*** .000

Activity status: Education 0.481*** .000 0.479*** .000

Activity status: Unemployment (ref.

cat.)

Social class: Lowest -0.264** .001 -0.264** .001

Social class: 2 0.049 .666 0.052 .650

Social class: 3 0.006 .936 0.007 .929

Social class: Highest (ref. cat.)

Identity (Minority) -0.569*** .000 -0.545*** .000

Political efficacy 0.131*** .000 0.131*** .000

Political involvement 0.321*** .000 0.323*** .000

Contextual predictors:

Welfare State Type: Post-socialist 0.825** .004

Welfare State Type: Nordic 1.316* .015

Welfare State Type: Conservative 1.174*** .000

Welfare State Type: Mediterranean 0.819* .030

Welfare State Type: Liberal (ref. cat.)

Cameral regime (bicameral) 0.520* .030

Representation regime: Proportional -0.098 .487

Representation

MYPLACE August 2014


regime:Majoritarian(ref. cat.)

Representation regime:Mixed (ref. cat.)

***p<0.001, **p<0.01, *p<0.05

The data also corroborates that social class has an impact on voting behaviour. Respondents from

lower social classes see their log odds of voting decrease 0.264 times when compared with the

higher social class. The same is true for minority groups: belonging to a minority group reduces

0.545 log odds of voting, which suggests that they are about half as likely to vote when compared

to the general population.

Regarding policy efficacy and political involvement, the results indicate that both increase the log

odds of voting in national elections. However this impact is greater when analysing the coefficient

associated with political involvement. The increase of 1 unit in the level of political involvement

increases 0.323 log odds of voting in national elections. With respect to the contextual variables it

is possible to observe that, as expected, the welfare state regime has an impact on the voting

behaviour of its citizens. Countries with the Nordic type of welfare state have more favourable

voting behaviours. Living in a country with this type of welfare state increases 1.316 log odds of

voting when compared with states with liberal regimes. When other welfare state regimes’ results

are observed we may conclude that liberal regimes are those who most contribute to the

abstention in national elections.

Regarding cameral system, as originally expected bicameral system favours turnout rates.

Bicameral regimes increases 0.52 log odds of voting when compared with countries with

unicameral systems. Type of parliament representation does not have impact on vote behaviour,

according to our results.

6.4.2 Model 2: Voting behaviour – Local elections

In order to analyse the voting behaviour in local elections we have adopted the same strategy that

was used in the analysis of national elections, once the data was organised in the same way,

according to a hierarchical structure. Regarding individual predictor variables, we have used the

variables listed in previous analysis. This will allow us to observe if the same variables have

different impacts when considering different types of elections.

With respect to the contextual variables it was necessary to introduce some changes to adapt the

analysis to local elections. In this case we maintained the hypothesis that welfare regimes, cameral

system, parliamentary representation have an impact on voting behaviour, but we have added

two more variables: the population size and the type of mayor elections (direct or indirect).

According to literature, in small localities it is possible to observe higher rates of voter turnout as it

allows greater proximity between the candidate and the population. In respect to the way mayors

MYPLACE August 2014


are elected, it is expected that direct elections encourage participation because people have a

higher perception of the impact of their vote.

After performing a multilevel analysis the results from the empty model suggest that only 7% of

variance lies between localities (ICC= 0.07). However this amount of unexplained variance is still

statistically significant (z=3.576; p<.01), allowing us to proceed with the multilevel analysis. The

introduction of individual variables reduced the variance between localities to 5.8% (ICC=0.058),

but still statistically significant (z=3.460; p<.01). In this way, we progress to a third model

introducing contextual variables. The introduction of these variables allowed the reduction of the

unexplained variance to 3.3% (ICC=0.03). Therefore, the first conclusion that we can pull out from

the comparison of national and local election is that localities introduce more variance in national

elections that in local elections. In other words, there are more differences between localities

when analysing national elections than local elections.

Table 6.4.2.1 Variance components of regression models explaining vote behaviour in local

elections

Empty model

(random

intercept

only)

With

individual

level

explanatory

var’s

With country

level

explanatory

var’s

Vote in local elections σ (Location level) 0.268 0.204 0.112

Z

3.576

(.000***)

3.460

(.000***) 2.622 (.000***)

AIC 33123.487 33993.619 34055.130

BIC 33130.434 34000.564 34062.073

-2 log pseudo log

likelihood 33121.487 33991.618 34053.129

***p<0.001, **p<0.01, *p<0.05

When we analyse the contribution of each predictor variable to voting behaviour in local elections,

in a more global analysis, it is possible to find out that the predictor variables, in general, follow

the same pattern identified in national elections. The main difference lies on the role of gender.

On the opposite of what we have found in national elections, gender does not contribute to

predict the voting behaviour in local elections. Regarding age, as observed in national elections,

there is positive impact of this variable on voting behaviour. The increase of 1 year in age will

increase 0.076 log odds of voting in local elections. Similar results are obtained regarding religion.

The increase of 1 unit on religion level will increase 0.041 log odds of voting. Social class also has

an impact on turnout rates, since it is between higher social classes that we have higher chances

to find respondents, which have voted in the last local elections. Similar results are, obtained

through education: the lower the level of education the lower log odds of voting. As seen before,

unemployed respondents are those who have shown lower log odds of voting when compared

MYPLACE August 2014


with any other employment status. Being employed or in education multiplies 0.344 and 0.321

respectively the times of log odds of voting when comparing to unemployed respondents. Also,

minorities report a significant decrease on log odds of voting. According to our data, belonging to

a minority decreases 0.429 log odds of voting in local elections.

Table 6.4.2.2 Parameter estimates of regression models explaining vote behaviour in local

elections (n2=30, n1=7687)

Empty model

(random intercept

only)

With individual level

explanatory var’s

With country level

explanatory var’s

SE SE SE


Intercept 0.199* .038 -2.392*** .000 -3.302*** .000

Age 0.077*** .000 0.076*** .000

Gender (male) 0.052 .366 0.052 .362

Education: ISCED 0-2 -0.861*** .000 -0.863*** .000

Education: ISCED 3-4 -0.323 .088 -0.321 .085

Education: ISCED 5-6 -0.106 .533 -0.097 .564


Religion 0.040** .001 0.041** .001

Activity status: Employment 0.349*** .000 0.344*** .000

Activity status: Education 0.331** .002 0.321** .003

Activity status: Unemployment (ref. cat.)


Social class: 2 -0.211* .028 -0.202** .037

Social class: 3 -0.093 .391 -0.091 .406


Identity (Minority) -0.467** .002 -0.429** .000




Welfare State Type: Post-socialist 0,93*** .000

Welfare State Type: Nordic 1,103*** .000

Welfare State Type: Conservative 1,208*** .000

Welfare State Type: Mediterranean 1,455*** .000

Welfare State Type: Liberal (ref. cat.)

N population -0,000 .085

Major election (direct) 0,117 .604

Representation regime: Majority 0,195 .270

Representation regime: Mixed -

Proportional and Majority -0,400*

.025

MYPLACE August 2014


Representation regime: Mixed electoral

system: individual constituency and party

list (compensation list).

0,736**

.001

Representation regime: Mixed-member

proportional representation 0,804

.092

Representation regime: Proportional(ref.

cat.)

***p<0.001, **p<0.01, *p<0.05

The perception that ordinary citizens may have an influence on political decisions (political

efficacy) as a positive impact on vote behaviour. The increase of 1 unit on political efficacy level

increases 0.105 log odds of voting. It is possible to find similar results regarding political

involvement. Citizens who participate more in political activities show also higher log odds of

voting. In respect to contextual predictors, it is possible to conclude that living in a country with a

liberal welfare regime reduces the log odds of voting in local elections.

The opposite of what we expected, in respect of population size and the way the mayor is elected

did not have an influence on turnout rates. However the parliamentary regime does: Mixed -

proportional and majority regimes reduces the log odds of voting in local elections when

compared with proportional regimes, but representations that mixed individual constituency and

party lists (compensation list) increases 0.736 log odds of voting when compared with reference

category - proportional representations.

6.4.3 Model 3: Closeness with political parties

Given the importance of parties in the political system and the proximity of citizens to parties as a

form of socialisation and political intervention, we intend to investigate some factors that

influence the closeness of young people to parties. This analysis is particularly relevant in the

context of the changes that are occurring in the way parties are relating to their members and

with the general population all across Europe.

To address this issue, in addition to socio-demographic variables, we have chosen a set of

variables that have been shown to have an impact on the relationship between citizen and states.

In this sense, three variables were chosen: political participation, political efficacy and trust in

political systems. Although it is predictable that these three variables have positive impacts on the

identification with a particular political party, it is not clear to what extent this influence takes

place and the variable with more prominence in this influence.

Regarding contextual variables, it was possible to observe that these have a different impact on

voting behaviour, including the characteristics of the current political system. We believe that

these variables will also impact on the closeness of young people to parties. A multilevel analysis

MYPLACE August 2014


will determine the way individual and contextual variables influence party identification of the

young people who participated in this study.

The results from the empty model reveal that 10.4% of variance in closeness to parties lies

between localities (ICC=0.104). This difference between localities is statistically significant

(Z=3.686; p<.001). The introduction in the model of individual variables allowed a significant

reduction on this variance from 10.4% to 3.7% (ICC=0.037), but still statistically significant

(z=3.455; p<.001), allowing the introduction of contextual variables in the model. However, in this

case the third model revealed a weak influence of context variables closeness to parties between

on variance. Since, in this model the variance that lies between localities reduced less than 1%

(ICC=0.03).

Table 6.4.3.1. Variance components of regression models explaining closeness with political

parties

Empty model

(random

intercept

only)

With

individual

level

explanatory

var’s

With country

level

explanatory

var’s

Closeness with

political parties σ (Location level) 0.384 0.138 0.112

Z

3.686

(.000***)

3.445

(.000***) 3.030 (.002)**

AIC 56480.528 57917.956 57947.012

BIC 56488.000 57925.427 57954.483

-2 log pseudo

loglikelihood 56478.527 57915.956 57945.012

***p<0.001, **p<0.01, *p<0.05

The analysis of coefficients of the third model allow us to conclude that, in parallel with what has

been observed in voting behaviour, age has a positive relation with closeness to political parties.

One year increase on age increases 0.064 log odds of being close to a party. It is also possible to

see that being male increases the log odds to be close to a party by 0.175 units. Conversely, to

what has been demonstrated in regards to voting behaviour, neither education nor employment

status or religion have an influence on the closeness to a political party. This is also true to ethnic

groups: belonging to an ethnic minority does not have an influence in respect to closeness to a

political party. However social class does. When we use the highest social class as reference

category the multilevel regression reveals that belonging to the lowest social class will decrease

0.209 log odds of being close to a political party. Regarding political efficacy, political involvement

and trust in political system, all three predict party identity. Political involvement is the variable

that most predicts the closeness to a party.

MYPLACE August 2014


Regarding contextual variables, we can conclude that welfare state regime does not influence

party identification. However the same is not true in respect to different countries’ political

systems. A bicameral regime decreases 0.308 log odds of being close to a political party and living

in a country with proportional representation parliaments will increase 0.358 log odds of being

close to a political party.

Table 6.4.3.2 Parameter estimates of regression models explaining closeness with political

parties (n2=30, n1=12997)

Empty model

(random intercept

only)

With individual level

explanatory var’s

With country level

explanatory var’s

SE SE SE


Intercept -0.343** .002 -3.101*** .000 -3.185*** .000

Age 0.065*** .000 0.064*** .000

Gender (male) 0.175*** .000 0.175*** .000

Education: ISCED 0-2 -0.269 .121 -0.279 .107

Education: ISCED 3-4 -0.104 .579 -0.113 .548

Education: ISCED 5-6 -0.125 .484 -0.131 .465


Religion 0.014 .218 0.016 .160

Activity status: Employment 0.056 .520 0.051 .567

Activity status: Education -0.038 .655 -0.045 .598

Activity status: Unemployment

(ref. cat.)


Social class: 2 -0.109 .207 -0.108 .207

Social class: 3 -0.116* .047 -0.117* .046


Identity (Minority) -0.134 .238 -0.135 .238



Trust in political system 0.064*** .000 0.062*** .000


Welfare State Type: Post-

socialist -0.108 .454

Welfare State Type: Nordic 0.395 .180

MYPLACE August 2014


Welfare State Type:

Conservative 0.226 .071

Welfare State Type:

Mediterranean 0.063 .765

Welfare State Type: Liberal (ref.

cat.)

Cameral regime (bicameral) -0.308* .041

Representation regime:

Proportional 0.358*** .000

Representation

regime:Majoritarian(ref. cat.)

Representation regime:Mixed

(ref. cat.)

***p<0.001, **p<0.01, *p<0.05, 13.8.2

6.4.5 Model 5: Non-electoral participation (Global indicator)

As it has been shown above, there is a substantial difference between locations in their level of

participation. By using multi-level modelling, we want to determine where this difference comes

from. Firstly, we construct a model for the global non-electoral participation model that is on a 0-

60 scale and contains information about the 20 political actions from the (Q16) of the MYPLACE

survey. The scale captures whether the respondent has or has not undertaken once, twice or three

or more times each of the 20 political activities.

Multi-level analysis allows the introduction of both individual and contextual variables in the same

model. For the individual level, we consider a set of variables trying to take into account the

different factors that the literature has defined as influential for youth participation. The model

contains basic socio-demographic attributes such as sex and age (treated as a continuous

variable). We have also included the activity status with a three-category variable (Working,

Studying and Unemployed or inactive) where ‘Working’ is the reference category. This variable

provides information about the influence of a crucial aspect for youth in the moment of their

transition to adulthood. The level of education is also in the model as a dichotomous variable,

considering people that have a post-secondary degree. In order to measure the social class of

young people, we have chosen the option of considering the labour situation and the level of

studies of parents. In so doing, we gain information about the social origin of young people in the

study. We use a 7-point interval scale where a higher level on the scale means a higher social class

position. Political attitudes are also considered in the model through two variables: Political

interest (a dummy variable), which is the most used indicator for psychological political

involvement, and dissatisfaction with the performance of democracy (also a dummy variable). We

have introduced this variable with the hypothesis that, in the context of crisis, the critique of the

system can work as a mobilization factor.

MYPLACE August 2014


For the contextual level, the data structure of the MYPLACE survey allows us to consider the

influence of location. In this case, we have introduced three contextual variables into the model.

The first one refers to the welfare regime (this variable has already been used and described in

precedent models). We have also considered the level of membership organisation in the location.

That is, the percentage of young people who stand and participate in an organisation. We know

that organisations are agents of participation, so we expect that in a context with a high level of

associations, young people will participate more. Finally, we have also included a variable in the

model to consider the contextual effect of the impact of the economic crisis. We would like to see

if societies that have been more affected by crisis, react by participating more or, on the contrary,

tend to be demobilised. To do so, we introduce the variation of youth unemployment for the

period 2007-2012.

The results of the variance in the different models show that 16.6% of the variance of the non-

electoral participation can be attributed to differences between locations. That means that the

rest, 83.4%, is explained by individual differences. Therefore, the first step is to introduce

individual variables into the model in order to discover how variance is reduced. In so doing, both

variances attributed to individuals and to locations are reduced. This is because different locations

also have a different composition of individuals (compositional effect). In this model, with

individual variables, the intra-class correlation shows that there is still a 13.4% variance that is

explained by contextual factors. Thus, we can also introduce contextual variables to improve the

model. When we introduce these variables at the level of locations, the intra-class correlation is

considerably reduced (1.7%).

Table 6.4.5.1. Variance components of regression models explaining non-electoral participation

Empty

model

(random

intercept

only)

With

individual

level

explanatory

var’s

With

country

level

explanatory

var’s

Non-electoral

participation

(global indicator)

σ (individual level) 27.,04 24.80 24.43

σ (location level) 5.38 3.83 0.41

Intra-class

correlation 0.166 0.134 0.017

The results in the table above show that the best model to explain non-electoral participation is

the one that includes individual and contextual variables. In the next table, we show the detailed

substantive results for this model.

Table 6.4.5.2. Parameter estimates of regression model explaining of non-electoral participation

(n1=12.951, n2=30)

MYPLACE August 2014


SE

Constant -6.585 *** 0.768


Activity status: Studying 0.684 *** 0.120

Activity status: Unemployed or inactive 0.015 0.146

Activity status: Working (ref. cat.) Ref. 0.000

Age 0.057 ** 0.019

Sex (female) -0.105 0.088

Post-Secondary education 0.570 *** 0.130

Social Class of origin 1.684 *** 0.135

Interest in politics 2.788 *** 0.093

Dissatisfaction with democracy 0.216 *** 0.020

Location level variables

Welfare R: Scandinavian 1.785 * 0.669

Welfare R: Continental 5.461 *** 0.877

Welfare R: Liberal 2.280 ** 0.623

Welfare R: Mediterranean -0.306 0.640

Welfare R: Post-Communist 1.186 * 0.458

Welfare R: Developing welfare 2.490 ** 0.674

Welfare R: Former URSS (ref. cat.) 0b 0.000

Level of organisation membership 0.055 *** 0.010

Youth Unemployment increase (2007-12) 0.139 ** 0.035

***p<0.001, **p<0.01, *p<0.05, ~p<0.1, two-tailed tests

The individual level results seem to confirm the expectable effect of the independent variables on

non-electoral participation. The only variable that does not have a significant effect on

participation is sex. The age variable shows a positive significant effect, so as young people get

older, they tend to participate more. Being a student also increases the probability of

participating. The educative environment and, in particular, the university act as a mobilizing

context (Caínzos, 2006). Another phenomena related to these results could be that when young

people start working, they have less time to be active in political actions. As we could expect,

having a higher education and coming from a family with a higher social class, also has a positive

effect on participation. Attitudes also have a role in explaining the propensity of being politically

active. Young people who are interested in politics tend to participate more. In addition, when we

look at participation as a whole, being critical of the performance of democracy also has a positive

effect on participation.

Looking at the results for contextual variables, welfare regime seems to have a role in the

participatory patterns of young people. Continental welfare regimes boost youth participation.

Locations with liberal and developing welfare regimes also show a greater propensity of having

higher levels of participation. The level of membership organisation shows a strong positive

MYPLACE August 2014


influence on the level of participation. Finally, the impact of the recent financial crisis and

measures that have increased youth unemployment, have had a negative effect on participation.

That means that, when looking at the global indicator of participation, the crisis has had a

demobilising effect on young people.

6.4.6 Models 6: Dimensions of participation

Individual and contextual factors can differently affect different dimensions of participation. As we

have seen, different dimensions of political action can be identified. These different dimensions

are different in their nature and probably will be influenced by different factors, both at the

individual and contextual level. In fact, once we look at the differences between locations in the

descriptive graphs, there are dimensions of political action that seem to be more influenced by

contextual factors than others. Both protest and private participation have a greater difference

between locations than public traditional participation. In order to confirm if the different

dimensions of participation are affected unequally by contextual factors, we need to look at the

intra-class correlation for the different dimensions of participation (in the empty model):

Table 6.4.6.1. Variance components of regression models explaining different dimensions of

political action

Private

individual

participation

Public

traditional

participation

Protest

action

Multi-level empty

model


σ (location level) 1.77 0.04 0.53

Intra-class

correlation 0.206 0.0197 0.173

The results show that while the variance attributed to locations in the case of private participation

and protest is high (20.6% and 17.3% respectively), in traditional public participation, the variance

for locations is very low (1.99%). This interesting result indicates that the use of institutional and

traditional forms of participation is more equal between young people all across Europe. This

similarity could be attributed to, the fact that it is a form of participation that is used by a very

small group of young people. This is also the case for protest and, in contrast, it does have a

significant location variation. Therefore, we can conclude that public traditional participation is a

more stable form of participation, compared to protest, and is less affected by the variations of

the context. This can be related to the fact that this dimension contains political activities that are

MYPLACE August 2014


oriented to institutions and, by definition, institutions are stable agents that do not change

considerably over time or across countries.

This small intra-class correlation tells us that it is not necessary to use a multi-level model to

explain public traditional participation. Thus, we will run a multi-level model for the remaining

participation indicators – private individual participation and protest action. In the next table, we

present the components of the variance in different models for these two dimensions of

participation:

Table 6.4.6.2. Variance components of regression models explaining private individual

participation and protest action

Empty

model

(random

intercept

only)

With

individual

level

explanatory

var’s

With

country

level

explanatory

var’s

Private individual

participation


σ (country level) 1.77 1.344978 0.245865

Intra-class

correlation 0.206 0.177 0.038

Protest action σ (individual level) 2.54 2.54 2.358420

σ (country level) 0.53 0.483857 0.151548

Intra-class

correlation 0.173 0.165 0.060

In both cases, the results indicate the need for a multi-level model with individual and location

level variables to understand variation in these kinds of participation. Once we introduce, these

two level variables, the intra-class correlation reduces substantially.

Table 6.4.6.3. Parameter estimates of regression models explaining Private individual

participation (n1=13.069, n2=30) and Protest action (n1=13.117, n2=30)

Private individual

participation

Protest action

SE SE

Constant -4.110 *** 0.605 -2.132 *** 0,459


Activity status: Studying 0.401 *** 0.061 0.082 ** 0.037

Activity status: Unemployed or inactive -0.079 0.074 0.105 ** 0.045

Activity status: Working (ref. cat.) Ref. 0.000 Ref. 0.000

Age 0.076 *** 0.010 -0.018 ** 0.006

Sex (female) 0.044 0.044 -0.056 ** 0.027

Post-Secondary education 0.347 *** 0.066 0.064 0.040

MYPLACE August 2014


Social Class of origin 0.887 *** 0.068 0.311 *** 0.042

Interest in politics 1.427 *** 0.047 0.480 *** 0.029

Dissatisfaction with democracy 0.099 *** 0.010 0.092 *** 0.006

Location level variables .

Welfare R: Scandinavian 0.760 0.593 1.323 * 0.461

Welfare R: Continental 2.821 ** 0.781 1.986 ** 0.608

Welfare R: Liberal 1.195 * 0.555 0.496 0.431

Welfare R: Mediterranean -0.069 0.571 -0.368 0.445

Welfare R: Post-Communist 0.606 0.407 0.534 0.317

Welfare R: Developing welfare 0.915 0.599 1.521 ** 0.466

Welfare R: Former URSS (ref. cat.) Ref. 0.000 Ref. 0.000

Level of organisation membership 0.027 ** 0.009 0.017 * 0.007

Youth Unemployment increase (2007-12) 0,043 0,031 0,089 ** 0,024

***p<0.001, **p<0.01, *p<0.05, ~p<0.1, two-tailed tests

In the case of private participation, results for individual variables are very similar to what we have

seen in the global non-electoral participation indicator. However, in protest action, we find some

differences. Firstly, age is still significant but it is negative for protest. Thus, as global participation

(and in private individual participation) increases with age, younger people are more active in

protest. It is also interesting to note that unemployed and inactive young people are more prone

to participate through protest actions than employed young people are. On the other hand, it

seems that the level of education is not crucial for protest as it is for other forms of participation.

When we look at location-level variables, there are still differences. In the case of welfare regime,

as in private individual participation, being in a continental or liberal welfare state appears to

increase the probability of participation. In the case of protest, the regimes that stimulate this

participation are the continental, the developing welfare regimes and Scandinavian. For both

forms of participation, the level of membership organisation in the location seems to be an

important variable. The more organisations there are in a locality, the more participative young

people are. Finally, the increase in youth unemployment seems to have an effect on levels of

protest but not for private participation. This indicates that in contexts where the impact of the

economic crisis has been harder, young people tend to choose more a more confrontational form

of participation.

MYPLACE August 2014


6.5 Summary


opportunity and 17% made a conscious decision not to vote. The remaining 11% reported that

they would have voted but were unable to. Locations with voting levels in excess of 80% were

Podsljeme, both Danish locations, Kuopio (Finland), both eastern German locations, Agenskalns

(Latvia) and Vic (Spain). By far the lowest level of voting was in Nuneaton (UK) where only 39% of

eligible voters chose to do so, with 43% of eligible voters deliberately choosing to stay away from

the polls.

More than half of respondents (59%) declared that they did not feel close to any political party.

In this chapter we analysed potential predictor variables of youth political activism. In order to

address this issue we opted to use multilevel regression analysis as MYPLACE data is organised in a

hierarchical structure. Individuals are nested in localities and localities are nested in countries. One

important dimension of political activism is voting behaviour. From the multilevel analysis

performed it was observed that it is in national elections that there is a higher level of variance

between localities, compared to local elections. In other words, contextual variables have more

impact in national elections than in local elections.

The results of variable coefficients allowed us to conclude that predictor variables follow very

similar patterns in the two types of elections. Older voters have higher log odds of voting. Religion

also has an influence on voting behaviour. Those who declared higher levels of religious identity

tend to demonstrate higher levels of participation in local and national elections.

As seen in previous studies, lower levels of education, belonging to a lower social class and

minority groups decrease the log odds of voting when compared to more privileged groups, such

as those with higher education, higher social class and integrated in the ethnic majority. Gender is

one of the few predictors that appears to behave differently with respect to national or local

elections. Males have higher log odds of voting in national elections than females, but this

difference is not statistically significant regarding local elections.

According to the literature on political efficacy, political involvement has a positive impact on

turnout rates. Our results support this statement. The increasing of political efficacy and political

involvement levels also increases the log odds of voting. Regarding contextual variables, we can

see that different political systems have different effects on voting behaviour. Liberal welfare state

regimes tend to discourage turnout. Bicameral regimes and proportional representation

parliaments increase the log odds of participation in elections.

We have also analysed closeness to political parties among young people. According to the data,

10% of the variance of youth closeness to political parties is explained by the differences between

MYPLACE August 2014


localities. A more detailed analysis shows a lower predictive ability of socio-demographic variables

such as employment status, educational attainment or ethnic belonging; none of these variables

has statistically significant capability to predict the closeness of the participants to a specific

political party. However, the same is not true of social class, it was observed, that belonging to the

higher social class increases the log odds of proximity to a political party.

Political efficacy, political involvement and trust in the political system also have an influence on

party identity; increasing levels of efficacy, political involvement and trust in the political system

result in higher log odds of feeling close to a political party. Regarding contextual variables only

cameral regimes and type of parliament representation influence the closeness to political parties.

Results suggest that living in countries with unicameral regimes and proportional representation

parliaments promote closeness to politics and decision making and, in this way, to political parties.

The MYPLACE survey allows for extended analysis of non-electoral political actions. This is crucial

when analysing young peoples’ participation since a number of published studies point to the

diversification of the political repertoires of young people (Soler-i-Martí, 2013). In addition, the

MYPLACE data permit the study of the impact of factors at the individual and the contextual level.

This is very interesting given the huge variation in the level of political activismacross Europe. In

order to capture diversity in forms of participation, we first have to identify different dimensions

in the participation repertoires of young Europeans. It is interesting to note that the first

dimension in a factorial analysis identifies a private and individual form of participation in which

the use of the Internet or political consumerism play important roles. This is the form of

participation more prevalent among young people. It is also the form of participation that is less

demanding in time and other costs such as exposure to the public since it usually occurs in the

private sphere. The other two dimensions of political participation that are identified capture

some more classical dimensions of political participation: traditional, and more institutional,

participation (that we have called public traditional participation); and protest actions (both legal

and non-legal). It is a minority of young people who carry out these two forms of political action.

Nevertheless, especially in the case of protest, there is great variation across countries and

locations.

Consequently, the focus of this chapter has been the explanation of differences across locations in

levels of participation. When we look at the global level of non-electoral participation, the

differences between the 30 locations in the study are huge. As an example, the two locations in

Hungary present a level of participation 25 times lower than some locations in eastern Germany or

Spain. Some of these differences are caused by what is called the compositional effect. That is that

the different locations have different compositions of individuals. However we have also analysed

the effect of contextual variables. We have seen that welfare regimes have an impact on

participation and that the level of membership in organisations improves the participation level of

young people in a location. We also wanted to see the effect of the economic crisis on the

participation patterns of young people. The results for the global indicator of non-electoral

MYPLACE August 2014


participation show that the impact of the crisis in youth unemployment has a negative effect.

However, it is interesting to highlight that when we analysed protest actions specifically, the

opposite effect is found. Therefore in locations that have experienced a greater impact of the

crisis, young people participate less in global terms but tend to participate more through protest

actions. In these contexts, where the dissatisfaction with institutions is higher, young people opt

for a more confrontational form of participation.

References

Blais, A., Gidengil, E., Nadeau, R. and Nevitte, N. (2001) Measuring Party Identification: Britain,

Canada, and the United States. Political Behavior, 23, 1, 5-22.

Dalton, R. (2007) The Good Citizen. How a Younger Generation is Reshaping American Politics.

Washington: CQ Press.

Education, Audiovisual and Culture Executive Agency (EACEA). (2013) EACEA 2010/03: Youth

Participation in Democratic Life. London School of Economics Final report. Retrieved April 15, 2014,

from

http://eacea.ec.europa.eu/youth/tools/documents/lse_study_on_youth_participation_2013.pdf.

Fieldhouse, E., Tranmer, M. and Russell, A. (2007) Something about young people or something

about elections? Electoral participation of young people in Europe: Evidence from a multilevel

analysis of the European Social Survey. European Journal of Political Research. 46: 797-822.

Franklin, M., Lyons, P. and Marsh, M. (2004) Generational Basis of Turnout Decline in Established

Democracies. Acta Politica 39 (2): 115-151.

Freire, André (2001) Modelos do Comportamento Eleitoral, Oeiras, Celta Editora.

Goerres, Achim. 2007. Why are Older People more Likely to Vote? The Impact of Ageing on

Electoral Turnout across Europe, British Journal of Politics and International Relations, 9/1: 90-121.

Henn, M., Weinstein, M. and Wring, D. (2002) A Generation Apart? Youth and Political

Participation in Britain. The British Journal of Politics and International Relations 4 (2): 167-192.

Hooghe, M., Kern, A. (2013) Party membership and closeness and the development of trust in

political institutions: An analysis of the European Social Survey, 2002-2010. Party Politics, 21-13: 1-

13.

Huggins, R. (2001) The transformation of the political audience. In B. Axford and R. Huggins (eds.),

New media and politics. London: Sage.

http://eacea.ec.europa.eu/youth/tools/documents/lse_study_on_youth_participation_2013.pdf

http://www3.interscience.wiley.com/cgi-bin/fulltext/117963588/PDFSTART



MYPLACE August 2014


Lister, M. (2007) Institutions, Inequality and Social Norms: Explaining Variations in Participation.

British Journal of Politics 9: 20-35.

Magalhães, Pedro. (2001) Desigualdade, desinteresse e desconfiança: a abstenção nas eleições

legislativas de 1999.Análise Social35 (157): 1079-1093.

Manning, N. (2010) Tensions in Young People's Conceptualisation and practice of Politics.

Sociological Research Online 15 (4).

Micheletti, M., Follesdal, A., and Stolle, D. Eds. (2003) Politics, products and markets: Exploring

political consumerism past and present. New Brunswick: Transaction Press.

Norris, P. (2004) Young People and political activism: From the politics of loyalties to the politics of

choice. Report for the Council of Europe Symposium: Young people and democratic institutions:

from disillusionment to participation. Strasbourg, 27-28th November 2003.

Nye, J. (1997) Why people don’t trust government. Cambridge, MA: Harvard University Press.

O'Toole, T. (2003) Engaging with Young People's Conceptions of the Political. Children's

Geographies 1(1): 71-90.

Park, A. (2000)‘The Generation Game’ in National Centre for Social Research British Social

Attitudes: Focusing on Diversity, London: Sage.

Pirie, M. and Worcestern, R. (1998) The Millennial Generation. London: MORI / Adam Smith

Institute.

Putnam, R. (1993) Making democracy work: Civic traditions in modern Italy. Princeton, NJ:

Princeton University Press.

Putnam, R. (2000) Bowling Alone: The Collapse and Revival of American Community. New York:

Simon and Schuster.

Soler-i-Martí, R. (2013) Democràcia, participació i joventut. Barcelona: Generalitat de Catalunya.

Stolle, D. and Hooghe, M. (2005) ‘Politics in the Supermarket: Political Consumerism as a Form of

Political Participation’ International Political Science Review 26 (3): 245-269.

Viegas, J. M., and Faria, S. (2003) A abstenção nas eleições legislativas de 2002. Retrieved April 15,

2014, from

http://www.ics.ul.pt/ceapp/conferencias/portugalavotos/viegasfaria/aabstencaoeleitoralde2002.

pdf.

Wattenberg, M. (2002) Where have all the voters gone? Cambridge, MA: Harvard University Press.

http://www.ics.ul.pt/ceapp/conferencias/portugalavotos/viegasfaria/aabstencaoeleitoralde2002.pdf

http://www.ics.ul.pt/ceapp/conferencias/portugalavotos/viegasfaria/aabstencaoeleitoralde2002.pdf

MYPLACE August 2014


Zukin, C., Keeter, S., Andolina, M., Jenkins, K. and Delli Carpini, M. X. (2006) A New Engagement?

Political Participation, Civic Life, and the Changing American Citizen. New York: Oxford University

Press.

MYPLACE August 2014


Chapter 7: Civic Engagement

Literature

Over the past 20 years, there has been considerable focus on youth civic engagement, mainly

because of concerns about the erosion of civic engagement in the younger generation, as has

been the case in the United States (Putnam 1995; 1997; 2000). In addition to the generational

‘problem’, however, there are also a number of other relevant issues to explore in order to

understand young people’s civic and political participation. Based on survey data on youth

engagement in political and civic associations, we aim to explore the socially conditioned

inequalities in civic and political participation and how structural factors influence young people’s

participation in the civic and political spheres.

Civic participation is often defined very broadly as individual and collective activities performed to

make a difference in society. Consequently, civic participation potentially covers many different

and sometimes quite diverse activities (Berger 2009, Ekman and Amnå 2012). In this chapter we

focus only on youth involvement in political and civic, and therefore we are mainly interested in

collective forms of organised formal participation.9 This kind of participation typically involves

citizens’ involvement in associations or organisations that pursue certain policy objectives or

interests through organised and collective activities vis a vis the political system or organise

certain civic activities.

In particular, associations like political parties and trade unions are important for the organisation

and representation of various social groups and political interests in society. They are seen as

important for democracy, not only because many of them function as mobilisers and channels of

political influence (Rosenstone and Hansen 2002, Warren 2001), but also because organisations

play a role as ‘schools of democracy’ (Tocqueville 1840). Seen from this perspective the engaged

members of associations develop a combination of social, associational, and political skills and

resources, such as self-confidence, associational and political knowledge and social contacts.

Political and civic associations play a significant role for the creation of local and national social

networks, and they can be seen as a link between citizens and the political system. Some have

very close connections to the established political systems, while others mainly are working at a

grass roots level, or in other important spheres of society such as religion and culture. In other

words, there are major differences between the various kinds of associations with regard to their

member base and forms of civic and political involvement and activities. During the last fifty years,

9 In their suggestion for a new typology of political participation and civic engagement Ekman & Amnå (2012), propose

two basic dimensions: The first dimension distinguishes between social involvement, civic engagement, formal political participation and (extra parliamentary) activism; and the second dimension distinguishes between individual and collective forms.

MYPLACE August 2014


especially political parties and labour unions have experienced declining member rates in most

parts of the western world, while associations and organisations connected to the new social

movements or traditional activities have experienced greater levels of public engagement,

especially among the younger generations. Despite these differences, we have in this chapter

chosen to summarise participation in an overall measure of membership in civic and political

associations.

In the following section, we first summarise the theoretical assumptions underlying the

hypotheses tested. In the analysis section, we first present an overview of how levels of

membership of civic and political associations vary across study locations and countries in the

MYPLACE study, and then we perform a comprehensive multilevel analysis that takes into account

the variation at the individual level and at the country level.

Explaining youth civic and political participation across locations – theory and hypotheses

When explaining why people do not participate in politics, there is a very simple way of expressing

it: “because they can’t, because they don’t want to, or because nobody asked” (Brady et

al.,1995:271). These three reasons, basically refer to the lack of resources (time, money and skills);

the absence of psychological engagement (political interest and concern with political issues); and

the absence of recruitment networks. Furthermore, one could also add that in some places they

not allowed to participate. A great deal of political participation research has examined how

participation corresponds with a number of demographic factors such as gender, age, and

resources related to education and social class. In line with this research we are also interested in

examining differences in young people’s civic participation across different social backgrounds;

however we are especially concerned with the question of how political participation varies across

the different research locations and countries in the MYPLACE study.

Explaining cross-country variation

Previous research in political participation has shown that in addition to individual variation there

is a considerable variation at the national level, indicating that cultural and institutional conditions

affect civic participation. One of the first major empirical studies addressing this issue was The

Civic Culture study by Almond and Verba (1989/1963). They explored the levels of civic

engagement in five different nations, using a cross-country survey, and found that at that time the

USA had considerably high levels of engagement. Later the more extensive World Values Study

(see e.g. Inglehart, 1997) confirmed the general picture of country variation in civic participation. A

common feature of these comparative approaches is that civic participation is seen as rooted in

different cultures and value systems (like materialist and post materialist values).

Schofer and Fourcade-Gourinchas (2001) present a ‘structuralist’ view on civic participation. They

analysed voluntary association membership in 32 countries, based on the World Values Surveys.

They suggest that two institutional distinctions account for much of the variation in organisational

MYPLACE August 2014


membership across countries: First, the distinction between Statist versus non-Statist (liberal)

societies, and second the distinction between cooperate versus non corporate societies. These

dimensions reflect the differences in state structure and political institutions, which encourage or

discourage the development of various types of organisations and civic participation in

organisations. An interesting finding in their study is that the level of statism has a particularly

strong negative impact on the overall level of membership in a country, and that the level of

corporateness has a positive impact (Schofer and Fourcade-Gourinchas, 2001).

Two more recent analyses contribute to this debate, with quite similar results in terms of the

variation at the country level. In a comparative analysis of the variety of social capital in welfare

states regimes, Kääriäinen and Lehtonen (2006), found that different forms of social capital are

connected to welfare states in different ways: Mediterranean welfare states have higher levels of

close familiar networks, while citizens in Scandinavian welfare states are more likely to engage in

voluntary associations. In an analysis of membership of political associations in 17 western

democracies, Morales (2009) also provides evidence for variation in political participation across

countries. Differences in levels of education, the socio-economic characteristics of the citizenry

alone, cannot explain country differences, and therefore Morales suggests that the political

context is of critical importance, and especially the existence and networks of politically active

organisations enabling mobilisation of citizens (Morales 2009:209).

In our analysis of young people’s membership of political associations we apply a modified version

of the theory of welfare state regimes, which was originally introduced by Esping-Andersen (1990).

Esping-Andersen considers welfare states to be complexes of institutions regulating the relations

between the state, the economy and the households. Historically welfare states have emerged at

different times and have followed different paths. Based on the degree of de-commodification and

the type of stratification, Esping-Andersen distinguishes between three ideal types of welfare

regimes: the liberal, the conservative and the social-democratic. To some extent, these regimes

also reflect the geopolitical map, and hence in the literature they are also called the Anglo-Saxon,

the Continental, and the Scandinavian. De-commodification is about the degree of government

regulation of the market, which is reflected in how different welfare services are offered.

Stratification refers to the scale of the social policy regulation in terms of offering benefits to

citizens. Compared to other welfare states, the Scandinavian welfare regime has a high degree of

regulation and a more comprehensive universalism, in the sense that citizens have equal access to

public services and benefits. Liberal Anglo-Saxon welfare states are characterised by a low degree

of de-commodification and a low degree of universalism, while conservative/continental welfare

states are placed between the first two. In the literature on welfare states, there has been much

debate on whether these distinctions are sufficiently accurate and nuanced. Certainly, there are

many differences within each group of welfare state types.

Later studies have suggested that the South European welfare states have special features, and

therefore should be added as a special type. They are characterised by a relatively low degree of

MYPLACE August 2014


de-commodification and a limited universality. The Catholic Church has historically played a major

role in the provision of social services and benefits (Arts and Gelissen 2002, Kääriäinen and

Lehtonen 2006). Furthermore, the new (re-)established states in Eastern Europe, has also

increased the interest in this group of states that despite large individual differences, have some

historical commonalities in their previous connection to the Soviet Union. Since the dissolution of

communist (or socialist) regimes, all these states went through a rapid and profound economic

transition. These states continue to be characterised by a certain degree of universalism, as the

economic transformation helped to reduce the degree of de-commodification.

As mentioned earlier we think the welfare state typologies may help to provide a useful analytical

categorisation, however, when theoretically considering what characteristics are essential for the

development of political associations and the level at political participation, it is also important to

note that it is not only a question of universalism and de-commodification. As pointed out by

Schofer and Fourcade-Gourinchas (2001) and Morales (2009) the relationship between the state

and civil society is of vital importance, especially to what extent the state supports and cooperates

with free political associations and organisations. Table 7 shows how we have categorised each of

the countries in the MYPLACE-study.

Table 7 Welfare state types

Welfare state typology Universalism De-

commodification

Statist

tradition +

Liberal

tradition -

Corporate

+

Non-

corporate

-

Countries in

MYPLACE-

survey

Scandinavian/

Social democratic

+ + (-) + Denmark,

Finland

Continental/Conservative - (+) + Western

Germany

Anglo-Saxon/Liberal - - - - UK

Mediterranean/ (Catholic) - - (+) - Spain,

Portugal

Greece

Post-communist (+) - + - Eastern

Germany,

Russia,

Estonia,

Latvia,

Croatia,

Hungary,

Slovenia,

Georgia

Hypotheses

MYPLACE August 2014


Based on previous comparative studies, our overall hypothesis regarding young people’s

membership of political associations in that we will find significant differences between groups of

welfare states. We also expect that participation in youth membership of political associations is

most pronounced in welfare states characterised by a high degree of universalism, de-

comodification and corporate relations between state and civil society.

This analysis of the MYPLACE data takes into account the country level differences by exploring to

what extend structural characteristics, related to welfare regime affects, the levels of civic

engagement among young people. In the analysis we include also a number of variables at the

individual level.

Gender and participation

Despite a movement towards greater gender equality in political participation and representation,

research still shows consistent differences in male and female political engagement. Women are

less interested in politics and are less politically active than men (Box-Steffensmeier et al., 2004;

Hayes and Bean, 1993; Howell and Day, 2000). Regarding affiliation with political associations,

research shows that women are slightly less likely than men to be affiliated with any organisation,

whether political or non-political. On average men are also affiliated with more voluntary

associations than women. Furthermore, men and women are involved in different kinds of

associations: In general, men are more likely to be involved in political and professional

associations and organisations, while women are more likely to be involved in religious and

educational organisations (Burns et al., 2001). We hypothesise that these patterns of gender

differences will also be present in the MYPLACE study.

Age and civic participation

Previous studies of civic engagement have noted that involvement is relative low in early

adulthood and increases with age. Jennings and Stokers’ (2004) longitudinal study of a 1965

cohort, across age 18-50, showed that engagement in various kinds of civic organisations was

highest in adolescence and midlife, but lowest during transition to adulthood. These participatory

patterns across the life cycle patterns are very much dependent on familial and institutional

boundaries. Civic participation in early adulthood is often related to school and education

activities, while exposure to more policy-oriented organisations, such as trade unions and political

parties, typically develops when people have settled into careers and adult roles such as,

parenting, home ownership (Kinder, 2006).

As the MYPLACE survey only involves young people age 16-25, we will not be able to analyse

participation in political organisations across the life span. However, we still expect to find an age

effect. The years of early adulthood are in many ways a turbulent phase of life, which is

particularly marked by social instability in terms of attachment to family and local community.

During these years many young people leave, their family home due to their choice of education,

jobs etc. Studies of young people’s voting patterns have shown that voter turnout is actually

MYPLACE August 2014


higher for the youngest voters, still living with their parents, than for those who have just left their

family home.

Social class, education and participation

Political participation is characterised by great social inequalities. People with higher incomes,

stable life situations, and especially higher education are typically more likely to participate in

politics (Milbrath and Goel, 1977, Nie and Verba, 1987). The positive correlation between levels of

education and political participation is among the best documented findings in political research

(Nie et al., 1996.).This also applies for participation in social and political organisations

(Rosenstone and Hansen, 2002). Part of the explanation for this is that social inequality often is

rooted in family background. Well-educated parents are typically more inclined to engage in social

and political life and raise their children to participate in society. Research into political

socialisation shows that children growing up in such families, are more likely to acquire knowledge

about politics, and a greater degree of political confidence.

The MYPLACE survey includes information about respondents’ employment status and level of

education. Based on previous research on inequality and political participation, we also expect

that education and employment will have a significant impact on the level of participation.

Motivation, knowledge, trust and religiosity

Besides the contextual and the socio-economic variables, we have also included variables

measuring political motivation, trust and religiosity. We expect that levels of trust in political

institutions has a positive impact on memberships of political associations, however this

relationship may vary according to the type of welfare state. We expect that positive relationships

mainly exist in the Scandinavian type of welfare state. In the MYPLACE survey respondents were

asked about their level of trust towards to different societal institutions and organisations. Index

for political trust was constructed from four different items.10

7.1 Civic Engagement Index (Q18)

Our dependent variable is ‘membership of political associations’. We have chosen to

establish a composite measure of the number of memberships. Based on the concept of

‘bridging social capital’ (Putnam, 1993), we think this to a certain degree reflects the extent

to which respondents in the survey are likely to be organised in different (although

sometimes overlapping) political associations. For the analysis, we constructed a

10

We use questions from Q7. Index consist the following items: trust in PM (Q7_3), trust in EC (Q7_7), trust in local parliament (Q7_8) and trust in parties (Q7_13). Trust in European commission is included in the index, because we find that supranational EU-level politics is an inseparable part of the local politics as well. Political trust index ranges from 0 - 40. Higher score means greater trust. A Cronbach Alpha value of 0.86 indicates that this four item scale has excellent reliability as a composite measure.

MYPLACE August 2014


membership index, including 11 out of 15 organisations mentioned in the survey

questionnaire. The organisations included are: ‘political parties or youth sections of political

parties’; ‘religious or church organisations or religious youth organisations’; ‘trade unions or

youth organisations of trade unions’; ‘national or local youth parliaments’; ‘environmental

organisations’; ‘animal welfare groups’; ‘peace organisations’; ‘human rights organisations’;

‘national or cultural organisations (for instance diaspora or ethnic minority organisations)’;

‘women’s organisations’ ; and ‘anti-globalisation organisations or movements’. Thus, our

dependent variable refers to the number of political organisations each of our respondents is

joining. The distribution of the responses is zero-inflated and provides an eleven-point scale

ranging from 0 to 11, where high values indicate high civic engagement.

As shown below, there is a considerable variation in membership of political organisations across

the different MYPLACE locations.

MYPLACE August 2014


Q18: Membership of Political Associations


The distribution of the responses is zero-inflated and provides an eleven-

point scale ranging from 0 to 11, where high values indicate high civic

engagement.

Graph 7.1: Mean Membership of Political Associations by location

Narrative

The overall mean for civic engagement is 0.25 (n=16,597, sd=0.716). This

varies by location, with the highest mean score at Odense Center (Denmark)

with 0.98, and the lowest mean score at Ozd (Hungary) with 0.04.

There are quite clear clusters of locations; with locations in Denmark,

eastern Germany, and western Germany, all at the top of the table, with

higher levels of civic engagement. Post-socialist countries, including

Hungary, Slovakia, Georgia and Latvia, are clustered in the bottom of the

table, with lower levels of membership in political organisations. In general,

civic participation is at a low level and most differences within countries are

relatively small. Also, countries with the very low levels of civic engagement

(under 0.30) have small differences between city and rural areas.

When compared with previous studies of memberships of political

associations, we find quite strong similarities in the way in which countries

are ranked in relation to each other (see Schofer and Fourcade-Gourinchas,

2001 and Morales, 2007).

MYPLACE August 2014



As the next step of the analysis, we performed multilevel analysis in order to estimate the

overall individual and country variation in civic participation, across the different European

research locations. As individual level variables, we have included gender, age, social class,

minority/majority status, education, employment, religiosity, political knowledge and

political interest, and political trust.

Model: Level of civic engagement

Table 7.2.1 Variance components of regression models explaining civic engagement

Empty

model

(random

intercept

only)

With

individual

level

explanatory

var’s

With country

level

explanatory

variable

σ (individual level) 1 1 1

Civic Engagement σ (country level) 0.645 0.506 0.118

-2 log pseudo likelihood 85761.757 84646.070 84433.203

The analysis of variance components expresses that 84 percent of the variance in

membership engagement can be explained by differences between individuals; whereas 16

percent of the variance can be explained by localities (Intra-class correlation (ICC) = 0.16). In

order to explain variation at the level of localities, a multilevel analysis (Generalised Linear

Mixed Model) was run.

By adding individual level variables to the null model, intercept variance is reduced by 22

percent in order to explain differences between localities in civic engagement scores. When

the contextual variable (‘welfare state regime’) was added, it explained a significant

proportion of the differences as intercept variance reduced even 77 percent.

MYPLACE August 2014


Table 7.2.2 Determinants of civic engagement

n2=30, n1=15,058

Model 1

b SE


Women Ref. Ref.

Men -0.161 *** (0.040)

Age 0.046 *** (0.012)

High Social Class - 3 Ref. Ref.

Class 2 -0.199 ** (0.075)

Class 1 -0.087 (0.062)

Low Social Class - 0 -0.164 * (0.067)

Majority Ref. Ref.

Minority 0.148 * 0.062

Other Ref. Ref.

Employed 0.205 0.108

In education 0.220 * 0.103

Unemployed 0.296 0.161

Religious Ref. Ref.

Not religious -0.265 ** 0.092

Political knowledge 0.129 *** 0.032

Not at all interested in politics Ref. Ref.

Not very interested in politics 0.192 0.104

Quite interested in politics 0.521 *** 0.101

Very interested in politics 0.942 *** 0.032

Political trust 0.003 0.015

Contextual level predictors

Liberal Ref. Ref.

Mediterranean -0.711 *** 0.188

Conservative 0.315 0.091

Nordic 0.747 *** 0.204

Post-socialist -0.876 *** 0.203

Intercept -2.835 *** 0.366

***p<.001, **p<.01, *p<.05, ~p<.1, two-tailed tests.

Starting with the individual level variables, we see from the full model in table 7.2.2, that

age has a positive impact on participation, however somewhat surprisingly the statistical

effect is quite weak. Another surprising finding is that women seem to have more political

memberships than men do (p<.001). As one might expect, civic engagement is also related

MYPLACE August 2014


to social class status. Roughly, the higher the social class that respondents belong to, the

more likely they are to participate. Unexpectedly, young people who are unemployed are

more likely to be members, than students or people who are employed. However, the

differences between groups are not statistically significant.

Civic engagement varies only a little between ethnic majority and minority groups.

However, we do find that young people belonging to minority groups tend to have slightly

higher level of memberships. The reason for this is presumably that the associational

memberships are related to specific political minority interests and activities. The young

people who are religious are more likely to be members of political associations than the

non-religious group. The difference is statistically significant (p<.01).

As expected, young people who are interested in politics are also more likely to be members

of political associations. Respondents who consider themselves to be very interested in

politics are considerably more active than the respondents with less political interest. In

addition, the more young people have political knowledge, the more they participate

(p<.001). It is evident that the causal relationship between these variables and associational

membership may also follow the opposite direction, in the sense that the associational

membership will produce more politically interested and knowledgeable members.

Quite surprisingly, political trust is not related to civic engagement as the regression

coefficient of the political trust variable is at very low level. The different is not statistically

significant (p=0.836).

As shown in graph 7.1, membership of political associations, varies notably between

countries. A central question in this chapter was to what extent these differences persist,

when we take into account the variation on the individual level. As revealed in the bottom

of table 7.2.2, young people belonging to the Nordic welfare regime are clearly more likely

to be members of political associations than young people from other welfare regimes,

whereas Mediterranean and particularly post-socialist regime are the least active regimes

(p<.001 in all aforementioned regimes).

7.3 Summary

16.4 Civic Engagement: Summary

Consistent with previous studies, our study of young Europeans aged 16-25, has shown that

there is considerable variation across countries and the variation is greater than the

variation between different locations in each of the countries. Similarities with previous

MYPLACE August 2014


studies are also evident in terms of how the countries are ranked in relation to each other.

Young people from the Nordic countries (Denmark and Finland), clearly have the highest

levels of membership of political associations, while young people from Mediterranean

(Portugal, Spain and Greece) and post-communist countries (Hungary, Slovakia, Georgia and

Latvia) have the lowest levels.

Based on institutional theory, and consistent with previous studies, we suggest that these

differences can be explained best by the ways in which states cooperate and support

political organisations. In the Nordic countries, especially Denmark, many voluntary

associations and organisations receive direct and indirect public subsidies with the purpose

of developing associations and civic engagement. Part of the conditions for receiving

economic funding is that these associations must comply with basic rules regarding

democratic organisation and economy. This also applies to political parties. Also labour

unions are encouraged through the tax deduction that members receive for their

membership.

From an institutional perspective, it is also important how the relationship between the

state and unions has evolved historically. In former communist countries political

associations, which were in opposition to the established regime, weres banned and

citizens’ political participation was only accepted if it supported the communist regime.

In relation to individual level variables, a surprising finding was that young women seem to

have a higher rate of membership than their male counterparts. While previous studies

have indicated a trend towards decreasing gender inequality, our research provides

empirical evidence that young women are about to overtake men in terms of their

membership of political organisations.

References

Adler, R. P. and Goggin, J. (2005) What Do We Mean By “Civic Engagement”? Journal of

Transformative Education 3, 3: 236–53.

Almond, G. A. and Verba, S. (1989). The Civic Culture: Political Attitudes and Democracy in

Five Nations. Sage Publications, Inc.

Arts, W. I. L. and Gelissen, J. (2002) Three Worlds of Welfare Capitalism or More? A State-of-

the-Art Report. Journal of European Social Policy 12, 2: 137–58.

Box-Steffensmeier, J. M., De Boef, S., and Lin, T. (2004). The dynamics of the gender gap.

American Political Science Review, 98, 515–528.

MYPLACE August 2014


Burns, N., Schlozman, K. L. and Verba, S. (2001). The private roots of public action: Gender,

quality, and political participation. Cambridge: Harvard University Press.

Berger, B. (2009) Political Theory, Political Science and the End of Civic Engagement.

Perspectives on Politics 7, 02: 335–50.

Howell, S. E., and Day, C. L. (2000) Complexities of the Gender Gap. Journal of Politics 62, 3:

858–874.

Ekman, J, and Amnå, E. (2012) Political Participation and Civic Engagement: Towards a New

Typology. Human Affairs 22, 3: 283–300.

Esping-Andersen, G. (1990.) The Three Worlds of Welfare Capitalism. Princeton, N.J:


Hayes, B. C. and Bean, C. S. (1993) Gender and Local Political Interest: Some International

Comparisons. Political Studies 41, 4: 672–82.

Inglehart, R. (1997) Modernization and Postmodernization. Princeton, New Jersey: Princeton

University Press.

Kinder, D. R. (2006) Politics and the Life Cycle. Science 312(5782): 1905–8.

Milbrath, L. W. and Goel, M. L. (1977) Political Participation: How and Why Do People Get

Involved in Politics? Chicago: Rand McNally and Co

Morales, L. and Geurts, P. (2007). ‘Associational Involvement’. In Citizenship and

Involvement in European Democracies: A Comparative Analysis, ed. van Deth, J.W.,

Montero, J.R. and Westholm, A. 135-157. Routledge.

Morales, L. (2009) Joining Political Organisations: Institutions, Mobilization, and

Participation in Western Democracies. Colchester: European Consortium for Political

Research Press.

Nie, N. H., and Verba, S. (1987) Participation in America: Political Democracy and Social

Equality. Chicago: University of Chicago Press.

Nie, N.H., Junn, J. and Stehlik-Barry, K. (1996) Education and Democratic Citizenship in

America. Chicago: University of Chicago Press.

Putnam, R. D., Leonardi, R, and Nanetti, R.Y. (1993) Making Democracy Work: Civic

Traditions in Modern Italy. Princeton University Press.

Putnam, R. D. (1995) Tuning In, Tuning Out: The Strange Disappearance of Social Capital in

America». PS: Political Science and Politics 28, 4: 664–83.

Putnam, R. D. (2001) Bowling Alone: The Collapse and Revival of American Community. New

York, NY: Touchstone Books by Simon and Schuster.

Rosenstone, S. J. and Hansen, J. M. (2002) Mobilization, Participation, and Democracy in

America. 1 edition. New York: Pearson.

MYPLACE August 2014


Schofer, E. and Fourcade-Gourinchas. M. (2001) The Structural Contexts of Civic

Engagement: Voluntary Association Membership in Comparative Perspective. American

Sociological Review 66, 6: 806–28.

Tocqueville, Alexis de (1994/1840) Democracy in America. New York: Everyman’s Library.

Adolph A. Knopf.

Warren, M. E. (2001) Democracy and Association. Princeton, New Jersey: Princeton

University Press.

MYPLACE August 2014


Chapter 8: Effective Political Action

It was important to find out how effectively, in respondents’ opinion, certain activities can

influence politics in the respective country. An 11-point scale was used to measure

perceived effectiveness of each of the following actions:

Being active in a political party (Q19_1),

Working in voluntary organisations and associations (Q19_2),

Personally contacting politicians (Q19_3),

Gaining publicity through media exposure (Q19_4),

Boycotting or boycotting certain products (Q19_5),

Participating in public demonstrations (Q19_6),

Signing petitions (Q19_7),

Participating in illegal activities (Q19_8),

Participating in violent protest activities (Q19_9),

Distributing political messages on social media sites (Q19_10),

Voting in elections (Q19_11).

Analysis of respondents’ answers proceeds in three steps.

First, we apply factor analysis to explore the possibility of reducing variables to any

underlying factors. This analysis suggests that the survey questions track onto two

underlying variables — legal or nonviolent political activity, and illegal or violent activity. We

therefore develop a composite scale for each variable, each of which has high reliability and

internal consistency.

Second, having reduced variables to the underlying scales, we analyse the scales by survey

location across all partners (total of 30 survey locations). We present descriptive statistics

for each scale, discuss clusters of locations with similar descriptive statistics, and discuss any

obvious disparities between two locations within the same country.

Third, we develop a multi-level regression model to explain individual attitudes toward the

effectiveness of political action. We focus analysis on the scale that captures attitudes on

the effectiveness of legal or nonviolent activity. With the model we are able to explore both

individual-level and contextual-level determinants of these attitudes.

Literature

Much of the recent literature on youth political participation emphasises an increasing

cynicism that is associated with declining rates of political participation. That is, structural

and institutional changes may lead to social dislocation which makes young people less

MYPLACE August 2014


confident that political activity is likely to be effective (see, for example, (Horvath and

Paolini, 2014) for an overview specific to survey research on European youth).

The predominant general explanations for political participation might be usefully

categorised into modernisation, political culture, and social capital approaches. The

modernisation approach typically holds that economic development causes changes in

economic and social organisation. These changes, including increased urbanisation,

industrialisation, education, and technological development may produce political

participation that is either constructive (Lipset, 1959; Epstein et al., 2006; Boix and Stokes,

2003) or non-constructive (Huntington, 1968; O’Donnell, 1973).

Cultural approaches build on structural explanations by arguing that economic development

operates on individual attitudes and beliefs to affect political participation. Typically, the

modernisation process encourages increased political participation through the

development of “modern” attitudes of diffuse interpersonal trust and resistance to

authority (Almond and Verba, 1963) or “postmodern” pursuits of goals beyond immediate

economic requirements (Inglehart, 1997).

Finally, social capital approaches draw on both structure and culture to explain patterns of

political participation. In the most prominent example, historically developed patterns of

interpersonal trust, diffuse social networks, and norms of reciprocity serve as solutions to

the collective action problem of political activity, resulting in greater government

accountability. This effective governance provides positive feedback, encouraging greater

political participation and creating virtuous cycles of citizen engagement and good

governance. Conversely, more closed forms of social organisation create vicious cycles of

disengagement and ineffective governance (Putnam et al., 1993).

While these general approaches to political participation do not deal directly with

individuals' perceptions of the effectiveness of specific types of political activity, their logic

might potentially explain these perceptions. Other studies have noted that survey questions

on perceptions of efficacy are unlikely to track perfectly onto actual participation (Horvath

and Paolini, 2014). However, factors that drive changes in participation are likely to have

some effect on how young people perceive the effectiveness of participation. Indeed,

Horvath and Paolini identify social dislocation resulting from supranational political

institutions and globalisation, shifts in political attitudes, and economic instability that have

led young people to lose faith in the effectiveness of traditional, institutionalised forms of

participation like political parties, in favour of more individualised, expressive forms of

participation. Similarly, Forbrig (2005) emphasises that frustration with traditional

democratic institutional channels has led young people to shift patterns of political

participation (see also Taft and Gordon, 2013). Other studies in this vein emphasise the

decline of political institutions (Duin, 1997), personal economic considerations (Winston,

MYPLACE August 2014


2013), and dislocation due to structural change (Edwards, 2009) as determinants of

attitudes toward political participation. This section draws on these approaches to develop a

multi-level model that explores how group-level indicators for globalisation, political

accountability, and welfare regimes interact with individual-level demographic variables to

affect youth attitudes toward the efficacy of political action.

8.1 Effectiveness of political activity scales (Q19)

Factor analysis of all Q19 survey items suggested two underlying latent variables—one

representing effectiveness of legal and nonviolent activity (Q19_1, Q19_2, Q19_3, Q19_4,

Q19_5, Q19_6, Q19_7, Q19_10, Q19_11) and another representing effectiveness of illegal

and/or violent activity (Q19_8 and Q19_9).11

The legal/nonviolent scale exhibited high reliability (Cronbach alpha = 0.884) with strong

internal consistency (location level Cronbach alphas ranging from 0.798 for Odense Center

in Denmark to 0.942 for Trnava in Slovakia). The resulting scale indicates the individual's

mean response on nine questions of the Q19 series — those concerned with the

effectiveness of political activity that is legal or nonviolent (‘political party activity’, ‘work in

voluntary organisations’, ‘contacting politicians’, ‘gaining media exposure’, ‘boycotts or

“buycotts”’, ‘demonstrations’, ‘petitions’, ‘social media activity’, and ‘voting’).

The illegal/violent activity scale also exhibited high reliability (Cronbach alpha = 0.852) with

strong internal consistency (location level Cronbach alphas ranging from 0.742 for Rostock

to 0.931 for Sopron). The resulting scale indicates the individual's mean response on two

questions of the Q19 series—those concerned with the effectiveness of political activity that

is illegal or violent (violent or illegal protest activities).

11

We also explored the possibility of an alternative set of three underlying factors: traditional institutional activity (including political parties, voluntary organisations, contacting politicians, traditional media, and voting), nontraditional or ad-hoc activity (including boycotts, demonstrations, petitions, and social media), and violent activity (violent or illegal protests). While factor analysis provided some support for these three latent scales, factor loadings were relatively low, despite high reliability tests. While these composites are theoretically interesting, we have decided to proceed with the stronger empirical argument, the two factor solution.

MYPLACE August 2014


Q19 Scale A: Effectiveness of Legal/Nonviolent Activity

Variable Description Variable is an 11-point scale ranging from 0.0 (low effectiveness) to 10.0

(high effectiveness). The distribution of scores on the scale is approximately

normal, although slightly skewed toward lower values. The distribution is

somewhat zero-inflated, with 235 respondents (1.4% of all respondents)

registering a score of zero (“not at all effective”). Respondents registering a

score of zero do not appear to be, distributed uniformly across locations. The

greatest percentage of these respondents were, located in Podslijeme

(Croatia) and Kupchino (Russia), each of which contained 10.6% (25

respondents each) of the respondents that registered zero on the scale.

Several locations, on the other hand, contained no respondents that

recorded a zero on the scale, including Odense East (Denmark), Jena

(eastern Germany), and Rostock (eastern Germany).

Narrative The overall mean for effectiveness of legal or nonviolent political activity is

5.26 (n=15,318, sd=1.94). This varies by location, with the highest mean

score in Telavi (Georgia) with 6.49, and the lowest mean score in Sopron


There is some clustering of locations; both locations in Denmark, Georgia,

eastern Germany and western Germany near the top, indicating higher levels

of belief in the efficacy of legal or nonviolent action. Central European

countries including Croatia, Slovakia, and Hungary are clustered toward the

bottom, indicating lower levels.

Most countries had similar mean score between the two locations in the

given country; there does not appear to be any large differences between

locations within each country.

Graph 8.1a: Mean effectiveness of legal/nonviolent activity by location

MYPLACE August 2014


Q19 Scale B: Effectiveness of Illegal/Violent Activity Variable Description

Variable is an 11-point scale ranging from 0.0 (low effectiveness) to 10.0 (high

effectiveness). The distribution of scores on the scale is vaguely normal, skewed toward

higher values, and heavily zero-inflated, with 3110 respondents (18.4%) registering a

score of zero. These respondents do not appear to be, uniformly distributed across

locations. Podslijeme (Croatia) contained the highest percentage of respondents that

registered a zero at 5.8% (180 respondents). Kuopio (Finland) contained their lowest

percentage at 1.6% (49 respondents).

Graph 8.1b: Mean effectiveness of illegal/violent activity by

location

Narrative

The overall mean for effectiveness of illegal or violent political activity is 3.21

(n=16,038, sd=2.60). This varies by location, with the highest mean score in Telavi

(Georgia) with 5.51, and the lowest mean score in Rostock (eastern Germany) with

2.33. There is some clustering of locations; both locations in Georgia, UK, and Latvia

near the top, indicating higher levels of belief in the efficacy of illegal or violent actions.

eastern and western Germany, Spain, and Denmark clustered toward the bottom,

indicating lower levels.

Most countries had a similar mean score between the two locations. Indeed, except

for the two Georgian locations, the entire sample exhibits relatively low variation. The

means of the two Georgian locations (Telavi and Kutaisi), though, are both statistically

significantly different at 90% confidence from the next highest location. At this stage,

we have developed two potential explanations for this difference. First, the 2003 Rose

Revolution and subsequent series of illegal or violent protest in Georgia may serve as

an “availability heuristic,” providing a successful example that leads respondents to

assess this strategy as effective. Second, respondents in countries with a historical

tradition of political protests may be more likely to perceive any given protest event as

a legitimate public demonstration. In contrast, respondents from countries lacking this

tradition, or which have historically defined public protest as illegal, may be more likely

to perceive any given protest event as illegal or violent. At this point, however, these

hypotheses are speculative, and require additional testing. Note also that this

difference exists for each of this scales constituent individual items (Q19_8 and Q19_9).

MYPLACE August 2014


8.2 Effective political action (Q19) - Constituent items

Respondents were asked to indicate on an 11-point scale, “How effectively do you think the

following actions can influence politics in your country?” with zero indicating “not at all

effective” and 10 indicating “very effective.” Respondents were, asked about 11 specific

types of political activity, including:

Q19_1: Being active in a political party

Q19_2: Working in voluntary organisations and associations

Q19_3: Personally contacting politicians

Q19_4: Gaining publicity through media exposure (TV/radio/newspapers)

Q19_5: Boycotting certain products

Q19_6: Participating in public demonstrations

Q19_7: Signing petitions

Q19_8: Participating in illegal protest activities

Q19_9: Participating in violent protest activities

Q19_10: Distributing political messages on social media sites

Q19_11: Voting in elections

Chart 8.2 displays the mean responses for each of the individual items (the composite scales

are included for comparison). Due to the large sample size, the mean responses for most

items are statistically significant. The exceptions (participation in political parties, voluntary

organisations, and public demonstrations) do not appear to be statistically significant from

each other. Voting in elections received the highest mean response for effectiveness, while

violent protest activities received the lowest rating for effectiveness.

Below, we present results for each of the items asked.

MYPLACE August 2014


Graph 8.2a: Mean Responses for Q19 Items (full survey sample)

MYPLACE August 2014


Q19_1: Effectiveness of Being Active in a Political Party


11-point scale measuring respondents’ assessments about

effectiveness of participating in activities of political parties.

Graph 8.2b: Mean effectiveness of activity in a political party

Narrative

The overall mean for Q19_1 is 5.41 (n=16,283, sd=2.67). Mean

responses vary by location with the highest mean score in

Telavi (Georgia) with 6.91, and the lowest mean score in

Pescenica (Croatia) with 4.03.

There is some clustering of locations, with locations in Georgia

and Denmark in the top third, indicating higher reported

beliefs in the effectiveness of party participation. Locations in

Croatia, Greece, and Hungary were clustered in the bottom

third, more sceptical of the effectiveness of party

participation.

Most countries had similar mean scores between the two

locations; a notable exception was Russia (Vyborg and

Kupchino), with a 1.51 point difference between two mean

scores.

MYPLACE August 2014


Q19_2: Effectiveness of Working in Voluntary Associations and Organisations



effectiveness of working in voluntary organisations or

associations.

Graph 8.2c: Mean effectiveness of working in voluntary associations and

organisations

Narrative


responses vary by location with the highest mean score in

Kuopio (Finland) with 6.39, and the lowest mean score in

Forstate & Jaunbuve (Latvia) with 3.56.

There is some clustering of locations, with locations in Spain

and eastern Germany in the top third, indicating higher beliefs

in the effectiveness of participation in voluntary organisations.

Locations in Latvia, Hungary, and Croatia were, clustered in

the bottom third, indicating less effectiveness.


locations; a notable exception was Russia (Vyborg and

Kupchino), with a 1.3 point difference between two mean

scores.

MYPLACE August 2014


19_3: Effectiveness of Personally Contacting Politicians



effectiveness of personally contacting politicians.

Graph 8.2d: Mean effectiveness of personally contacting politicians

Narrative


responses vary by location, with the highest mean score in

Odense Center (Denmark) with 5.60, and the lowest mean score

in Perscenica (Croatia) with 2.78.

There is some clustering of locations, with locations in Denmark

and Georgia in the top third, indicating higher beliefs in the

effectiveness of contacting politicians. Locations in Croatia,

Greece, and Hungary were, clustered in the bottom third,

indicating less effectiveness.

Some countries exhibited differences between locations with

regard to the effectiveness of contacting politicians. The most

dramatic examples were Russia (Vyborg and Kupchino) with a

1.5 point difference between two mean scores, and Portugal

(Barreiro and Lumiar) with a 1.3 point difference.

MYPLACE August 2014


Q19_4: Effectiveness of Gaining Publicity through Media Exposure (TV/radio/newspapers)



effectiveness of gaining publicity through traditional media

(TV / radio / newspapers).

Graph 8.2e: Mean effectiveness of gaining media exposure

Narrative

The overall mean for Q19_4 is 5.93 (n=16,459, sd=2.69).

Mean responses vary by location with the highest mean

score in Odense Center (Denmark) with 7.73, and the lowest

mean score in Argyroupouli (Greece) with 4.25.

There is strong clustering of locations, with locations in

Denmark, Georgia and eastern Germany in the top third,

indicating higher beliefs in the effectiveness of gaining media

exposure. Locations in Croatia, Greece, Slovakia and Hungary

were, clustered in the bottom third, indicating less

effectiveness.

All countries have similar mean scores between locations.

MYPLACE August 2014


Q19_5: Effectiveness of Boycotting or Buycotting certain products



effectiveness of boycotts or “buycotts” of certain products.

Graph 8.2f: Mean effectiveness of boycotts

Narrative



score in Argyroupouli (Greece) with 5.92, and the lowest

mean score in Podsijeme (Croatia) with 3.32.

There is some clustering of locations, with locations in

Greece, western Germany, and eastern Germany in the top

third, indicating higher beliefs in the effectiveness of

boycotts. Locations in Croatia, Hungary, and Russia were,

clustered in the bottom third, indicating less effectiveness.

All countries appear to have similar mean scores between

locations.

MYPLACE August 2014


Q19_6: Effectiveness of Participating in Public Demonstrations



effectiveness of public demonstrations.

Graph 8.2g: Mean effectiveness of participation in public demonstrations

Narrative



score in Sant Cugat (Spain) with 6.41, and the lowest mean

score in Sopron (Hungary) with 3.55.


Spain, Georgia and eastern Germany in the top third,

indicating higher beliefs in the effectiveness of

demonstrations. Locations in Croatia, Hungary, Slovakia and

Russia were, clustered in the bottom third, indicating less

effectiveness.

All countries appear to have similar mean scores between

locations.

MYPLACE August 2014


Q19_7: Effectiveness of Signing Petitions



effectiveness of signing petitions.

Graph 8.2h: Mean effectiveness of signing petititons

Narrative



score in Jena (eastern Germany) with 6.19, and the lowest

mean score in Sopron (Hungary) with 3.55.


eastern Germany in the top third, indicating higher beliefs in

the effectiveness of signing petitions. Locations in Hungary,

Russia, Estonia and Croatia were, clustered in the bottom

third, indicating less effectiveness.

Most countries have similar mean scores between locations.

One exception was Latvia (Agenskains and Forstate &

Jaunbuve) with a 1.25 difference in means between

locations.

MYPLACE August 2014


Q19_8: Effectiveness of Participating in Illegal Protest Activities



effectiveness of illegal protest activities.

Graph 8.2i: Mean effectiveness of illegal protest activities

Narrative



score in Telavi (Georgia) with 5.42, and the lowest mean

score in Bremerhaven (western Germany) with 2.84.


Georgia, UK and Greece in the top third, indicating higher

beliefs in the effectiveness of illegal protests. Locations in

western Germany and Croatia were, clustered in the bottom



MYPLACE August 2014


Q19_9: Effectiveness of Participation in Violent Protest Activities



effectiveness of violent protest activities.

Graph 8.2j: Mean effectiveness of violent protest activities

Narrative




score in Rostock (eastern Germany) with 1.69.


Georgia, UK and Latvia in the top third, indicating higher

beliefs in the effectiveness of violent protests. Locations in

western Germany, eastern Germany, Spain and Denmark

were, clustered in the bottom third, indicating less

effectiveness.


MYPLACE August 2014


Q19_10: Effectiveness of Distributing Political Messages on Social Media Sites



effectiveness of distributing messages on social media.

Graph 8.2k: Mean effectiveness of social media

Narrative




score in Rimavska Sobota (Slovakia) with 3.40.


Georgia and Denmark in the top third, indicating higher

beliefs in the effectiveness of social media. Locations in

Slovakia, Hungary and Croatia were, clustered in the bottom



Exceptions include Latvia (Agenskains and Forstate &

Jaunbuve) with a 1.26 point difference in means between

locations, Russia (Vyborg and Kupchino) with a 1.09 point

difference, and Slovakia (Trnava and Rimavska Sobota) with a

1.10 point difference.

MYPLACE August 2014


Q19_11: Effectiveness of Voting in Elections



effectiveness of voting in elections.

Graph 8.2l: Mean effectiveness of voting in elections

Narrative




score in Kupchino (Russia) with 4.72.


Georgia, Denmark and eastern Germany in the top third,

indicating higher beliefs in the effectiveness of voting.

Locations in Russia, Croatia and Hungary were, clustered in

the bottom third, indicating less effectiveness.


Exceptions include Latvia (Agenskains and Forstate &

Jaunbuve) with a 1.23 point difference in means between

locations, and Russia (Vyborg and Kupchino) with a 1.29 point

difference.

MYPLACE August 2014


8.3 Multi-Level Modelling

In this section, we analyse the determinants of young people’s beliefs in the effectiveness of

legal or nonviolent political activity. To account for the nested structure of MYPLACE survey

data, i.e. individuals nested in locations, which are nested within countries, the analysis

employs a multilevel regression model that tests the effects of both individual-level and

group-level variables.

The analysis proceeds in three stages. First, we provide a conceptual overview of the

dependent variable, and several independent variables advanced by the relevant literature

that may be, expected to affect young people’s attitudes toward the effectiveness of legal

or nonviolent political activity. Second, we confirm that adding contextual level variables in

a multilevel analysis is appropriate through the analysis of variance components of models

with individual and group-level indicators added successively. Third, we present the

substantive results of a linear multilevel regression model. Initial findings suggest that

several demographic variables have a statistically significant effect on attitudes toward the

effectiveness of legal or nonviolent political activity, although that effect is substantively

small to moderate. Among individual level variables, internet usage and political knowledge

exhibited the largest substantive effects. The only group-level variable with a statistically

and substantively significant effect on youth attitudes toward the effectiveness of legal or

nonviolent politically activity was welfare regime type, with post-communist welfare

regimes associated with substantively lower beliefs in the effectiveness of this activity

relative to the reference category of liberal welfare regimes.

Dependent Variable:

The dependent variable is a composite scale of respondents’ ratings of the effectiveness of a

series of nine specific types of legal and non-violent political activity. As discussed in the

descriptive analysis section above, the MYPLACE survey asked young respondents, “How

effectively, do you think, the following actions can influence politics in your country?” with

reference to 11 specific actions. Effectiveness of each action was, rated by, the respondents

on an 11-point scale, with code 0 meaning “not effective at all” and code 10 meaning “very

effective.” A factor analysis of the 11-item series suggested two latent variables, one of

which might be, called “legal or nonviolent activity.” This variable captures nine of the

original items, including activity in political parties, voluntary organisations, contacting

politicians, gaining media exposure, boycotts, public demonstrations, petitions, distributing

messages by social media, and voting (illegal protest activity and violent protest activity

tracked onto a separate latent variable, according to the factor analysis). Therefore, the

composite variable “legal or nonviolent political activity” is a scale aggregating responses on

the nine constituent items. The scale ranges from 0 to 90, with higher scores indicating

MYPLACE August 2014


stronger beliefs in the effectiveness of legal or nonviolent activity.12 This scale is,

approximately normally distributed, although somewhat zero-inflated. The mean for the

entire survey sample is 47.78 with a standard deviation of 17.21 (N=14,269).13

Due to the approximately normal distribution of the legal or nonviolent political activity

scale, we have opted to employ a linear mixed model for analysis. Although conscious of the

potential for violations of regression assumptions with a zero-inflated distribution, at this

point, a linear model seems to be the best fit for the data. While negative binomial or

Poisson models are usually employed to account for zero-inflation, these distributions are

usually generated by “counts” or discrete incidents within a geographical area or time

period, which, does not match the data-generating process behind ratings of the

effectiveness of political activity. While composite variables can often approximate a

“count” distribution, making negative binomial or Poisson models appropriate, the

distribution on our dependent variable most closely approximates a classic normal

distribution. Furthermore, the zero-inflation in the distribution is small to moderate, with

only 1.3% of the sample (189 individuals) recording a score of zero on the scale. Finally,

linear regression allows for more straightforward interpretation of estimates, which is

perhaps more useful for an initial analysis. Publications intended for peer-review, however,

should explore additional tests for model fit, including standard tests for violations of linear

regression assumptions.14

Independent variables:

As noted in the brief literature review above, several individual-level variables might be,

expected to affect young people’s attitudes toward the effectiveness of legal or nonviolent

political activity. According to the logic of the modernisation theory, for example, structural

economic changes drive changes in distribution of income, access to information and

education that lead to increased political activity. Exposure to a greater variety of people

from different social, economic, or ethnic groups, combined with greater access to

information and increased economic resources may lead to increased consciousness of

12

Note that the scale used for the multilevel model is different from the conceptually identical scale analysed in the descriptive section. In order to facilitate potential analysis with a variety of models, including negative binomial and log-linear models, the scale of average responses ranging from 0-10 in the first section was converted to an integer scale ranging from 0-90 in this section. 13

Includes only individuals included in the multilevel model. Individuals were filtered prior to analysis to drop any cases with missing values on any variable. Filtering at the beginning of the analysis ensures that models used for comparison are run on an identical set of observations. 14

We did conduct analysis using alternative distributions for all models described below, including negative binomial and log-linear models for all steps. A particular problem for this analysis was that these models indicated that practically no variance (0.000) in the dependent variable was explained by group-level variables. While possible, we thought this unlikely given the nested data structure, the selection process for locations, and obvious differences in location and country-level means. However, having most or all variance explained at the individual level is certainly possible. We also conducted a linear mixed model dropping all respondents that registered “0” on the legal or nonviolent political action scale. However, this analysis suggested no major differences in the significance, direction, or magnitude of effects.

MYPLACE August 2014


political interests, making political participation more likely (Lipset, 1959; Almond and

Verba, 1963; Epstein et al, 2006; Boix and Stokes, 2003; Huntington, 1968; O’Donnell, 1973;

Przeworski et al., 2000). Applied to our dependent variable, these changes may lead young

people to rate legal or nonviolent political activity as more effective, at least according to

those promoting the modernisation theory (Lipset, 1959; Almond and Verba, 1963; Boix and

Stokes, 2003). On the other hand, if these changes outstrip the capacity of political

institutions to channel new political demands (see, for example, Huntington, 1968;

O’Donnell, 1973), they may lead young people to rate legal or nonviolent political activity as

less effective. To test these hypotheses among the population in the MYPLACE survey, we

include individual-level variables for average daily use of the internet use for information on

politics or current events (Q3_2), parental social class (since most respondents have not

entered the work force yet, or had the opportunity to transition between social classes, we

use parental social class as a proxy for the respondents’ social class), and political

knowledge (Q28s15).

Political culture approaches to democratisation suggest another set of independent

variables that might affect political activity. For political culture explanations, economic

structural changes drive changes in attitudes that are more conducive to democracy. For

instance, several cultural explanations have suggested that the degree to which individuals

trust people outside of their kin, ethnic, or economic group facilitates effective democratic

governance (see, for example, Almond and Verba, 1963 and Putnam et al., 1993). Putnam et

al. (1993) is perhaps most specific about the mechanisms linking trust and governance by

arguing that interpersonal trust makes citizens more likely to participate constructively in

collective action. We might therefore expect that higher levels of interpersonal trust are

associated with higher beliefs in the effectiveness of legal or nonviolent political action. To

test this hypothesis among MYPLACE respondents, we employ the item for interpersonal

trust (Q29) as an individual-level independent variable.

A third group of literature more specific to the issue of beliefs about the effectiveness of

political action argues that cynicism with political participation results from

disenfranchisement associated with economic or political globalisation, or from

disillusionment with governance. Horvath and Polini (2014), for example, find that

supranational political institutions, economic globalisation and economic instability have

caused young people to lose faith in institutionalised participation like political parties.

Several other studies have emphasised that disillusionment with democratic governance

have led to shifts in how young people evaluate the effectiveness of political activity

(Forbrig, 2005; Taft and Gordon, 2013; Duin, 1997). Finally, economic dislocation and

instability, in particular extended adolescence due to an absence of economic opportunities,

15

Although country-specific variance was allowed in questions measuring respondents’ knowledge about current politics in their country (Q28_1, Q28_2 and Q28_3), these questions provide reliable and comparable measure of how well the respondents are informed about national politics.

MYPLACE August 2014


drive young people’s cynicism about political participation. According to this same logic, we

might expect members of historically marginalised groups, including women and ethnic and

racial minorities, to have more pessimistic views about the effectiveness of legal or

nonviolent political activity. We test these hypotheses among the MYPLACE respondents

using individual-level variables, including respondents’ employment status (Q61R2),16

gender (Q54), minority/majority status (variable “identity”); and group-level variables,

including the Index of Globalisation (GI) and the Voice and Accountability indicator from the

World Bank Worldwide Governance Indicators. We also include the harmonised welfare

regime type variable (Kaariainan and Lehtonen, 2006), under the logic that a highly

developed welfare regime has the potential to mitigate many of the problems associated

with political and economic dislocation. Finally, age and religiousness are included as

individual-level control variables.

Results

Variance Component Analysis

A first step in developing a multilevel model is to analyse the variance components of a

series of increasingly complex models to determine whether adding indicators at different

levels of aggregation reduces unexplained variance. First, we derive the variance

components for an empty- or null-model for individuals clustered within survey locations,

with a random intercept and no individual- or group-level variables. The variance

components of the null model suggest an intra-class correlation (ICC), or the correlation

between two randomly selected respondents from the same survey location, of 0.132.17 The

ICC for the null model suggests that 13.2% of variance in ratings of effectiveness of legal or

nonviolent political activity is explained by differences at the group level (conversely, 86.8%

of variance is explained by differences between individuals), suggesting that a multilevel

analysis is appropriate.

Next, we develop a fixed effects model using only individual-level variables to determine

whether these indicators reduce unexplained variance at either individual- or group-level.

The ICC of this model was 0.119, indicating that the inclusion of individual-level variables led

to 11.9% of variance explained by differences between locations. Indeed, inclusion of the

individual-level variables led to a 4% reduction in residual individual-level variation, and a

14.5% reduction in variance at the level of locations.

Finally, we develop a fixed effects model that includes the group-level variables in addition

to the individual-level indicators. The ICC for the third model was 0.063, suggesting that

16

Variable Q61R was recoded to change the reference level to “unemployed.” 17

The ICC is derived by calculating the proportion of location-level variance to the total variance (location-level plus individual-level variance).

MYPLACE August 2014


6.3% of remaining variation is explained by differences between locations. Inclusion of the

group-level variables led to practically no reduction (0.00004) in individual-level variance,

but a 50% reduction in group-level variance.

The results of the variance component analysis suggest that only a small to moderate

amount of variance in attitudes toward the effectiveness of legal or nonviolent political

action is explained by differences in locations. While including group-level variables reduce

the intercept variance, they reduce little unexplained variance at the individual level.

Furthermore, the estimates in the individual-level variables only model (not reported) are

very similar in significance, direction, and magnitude to the multilevel model that includes

contextual variables. Also, the variance component analysis using negative binomial and

loglinear models produced ICCs of virtually zero, which would also seem to indicate that

little individual-level variance can be explained by differences in locations.

Table 8.3.1 Variance Components of Regression Models for Perceptions of Effectiveness of

Legal or Nonviolent Political Action

Null Model (Random

Intercept Only)

Fixed Effects Model

with Individual-Level

Variables Only

Fixed Effects Model

with Individual- and

Group-Level Variables


Scale: Perception of

Effectiveness of Legal

or Nonviolent Political

Action

σ (group level) 39.499110 33.783464 16.823747

-loglikelihood 119929.613 119312.330 119291.883

ICC 0.132 0.119 0.063

Substantive Results

Turning to the substantive results of the multilevel model, we find tentative support for

some of our hypotheses. In particular, at the individual level, both usage of internet to get

information about politics and current affairs and political knowledge exhibited statistically

significant and substantively large effects on respondents’ perceptions of the effectiveness

of legal or nonviolent political activity. Specifically, the mean effectiveness rating by the

respondents who spent “no time at all” using the internet for information on politics or

current affairs was 5.12 points lower than that by the reference group—respondents who

spent more than an hour during an average day using the internet for news or political

information, holding other variables constant. Similarly, the mean effectiveness rating

among respondents in the two lowest political knowledge category was approximately 5

MYPLACE August 2014


points lower than the mean rating for the highest political knowledge category.

Respondents that answered all three political knowledge questions correctly indicated a

mean effectiveness of legal and nonviolent political activity that was about five points

higher than those that answered one or no question correctly, holding other variables

constant. These results lend some support to the traditional modernisation hypothesis—

that the development of technology makes information more accessible, in turn leading to

more constructive political engagement.

Several other individual-level variables exhibited a highly statistically significant effect on

perceptions of the effectiveness of legal or nonviolent political activity, but those effects

were substantively small. Males tended to have lower perceptions than females of the

effectiveness of this type of political activity, averaging 2.33 points lower on the 90 point

scale, holding other variables constant. Similarly, non-religious people rated legal or

nonviolent activity less effective on average than did religious people, though the effect was

similarly small. Likewise, members of the lowest socioeconomic class rated legal or

nonviolent activity less effective on average than did respondents of the highest social class,

while respondents that were employed or in education exhibited slightly more optimism

about the effectiveness of this type of activity relative to the reference group of

unemployed respondents. Both interpersonal trust and age exhibited a statistically

significant and negative effect on perceptions of effectiveness of legal/nonviolent political

activity, though these effects were substantively very small: -0.3 and -0.13 on a 90-point

scale, respectively. The remaining individual-level variable, “identity”, was not statistically

significant.

Taken as a whole, these results provide little support for strictly economic or dislocation

hypotheses. Members of high socioeconomic class and respondents that were employed or

in education, as expected, were more likely to view legal or nonviolent political activity as

effective, but only marginally so. There is no evidence in these results that historically

marginalised groups are less optimistic about the effectiveness of this type of activity than

majority-group respondents. Finally, contrary to the expectations of political culture

approaches to civic activism, there is no evidence that higher ratings of interpersonal trust

are associated with perceptions of greater effectiveness of constructive political activity.

Among the group-level variables, only the “post-socialist” welfare regime group and the

Index of Globalisation were statistically significant. The effect of economic, cultural, and

political globalisation on perceptions of effectiveness of legal or nonviolent activity was

negative, but substantively small, -0.38 on a 90-point scale. The indicator for Voice and

Accountability, on the other hand, did not exhibit a statistically significant effect of

perceptions of effectiveness. These results, therefore, provide little support for the

hypotheses that supranational political, cultural, and economic changes cause dislocation

that makes young people more cynical about constructive political participation. Similarly,

MYPLACE August 2014


there is no evidence that effective governance, as measured by the Voice and Accountability

indicator, has any effect on young people’s attitudes toward the effectiveness of legal or

nonviolent political activity.

On the other hand, the indicator for the “post-socialist” welfare regime group was both

statistically significant and substantively large. Respondents in “post-socialist” welfare

regimes rated the effectiveness of legal or nonviolent activity a mean of 11.29 fewer points

than did respondents in the reference group, “liberal” welfare regimes. The other three

welfare regime categories did not exhibit a statistically significant effect. Taken at face

value, this result seems to contradict the hypothesis that liberal welfare regimes create

cynicism among youth because they do not mediate the dislocation of structural economic

changes. However, it is worth noting the only country in the sample categorised as a

“liberal” welfare regime is the UK, meaning that the result should be, interpreted as

respondents in post-socialist countries rate the effectiveness of legal or nonviolent political

activity lower on average than do respondents in the UK. This difference is likely due to

other contextual differences between these groups of countries, rather than to differences

in the welfare regime.

Table 8.3.2 Determinants of attitudes toward effectiveness of legal or nonviolent political

activity; n2=30, n1=14,269 Model 1

Individual-level predictors b Sig. SE

Use of internet18

: no time at all -5.178 *** (0.440)

Use of internet: less than ½ hour -2.006 *** (0.394)

Use of internet: ½ hour to 1 hour -0.153 (0.420)

Use of internet: more than 1 hour Ref.

Gender (male) -2.331 *** (0.270)

Gender (female) Ref.

Age -0.133 * (0.056)

Employed 1.242 * (0.507)

In education 1.781 *** (0.488)

Other -0.240 (0.696)

Unemployed Ref.

Class 019

(lowest socioeconomic class) -1.539 *** (0.367)

Class 1 -0.469 (0.485)

Class 2 -1.078 * (0.481)

Class 3 (highest socioeconomic class) Ref.

Identity 020

0.055 (0.400)

18

“On an average day, how much time do you spend keeping yourself informed about politics and current affairs using the following media: the internet?” 19

Harmonized variable computed based on father’s education (Q73), mother’s education (Q80), father’s

occupation (Q76), and mother’s occupation (Q80). 20

Identity 1=respondent in majority in ethnicity, citizenship, nationality, and country of birth categories,

Identity 0=all others; Harmonized variable computed based on original variables Q22, Q56, Q57, Q58.

MYPLACE August 2014


Identity 1 Ref.

Not Religious -1.607 *** (0.340)

Religious Ref.

Political Knowledge 021

(lowest political knowledge) -5.333 *** (0.522)

Political Knowledge 1 -4.836 *** (0.410)

Political Knowledge 2 -2.109 *** (0.351)

Political Knowledge 3 (highest political knowledge) Ref.

Social trust22

-0.301 *** (0.059)


Liberal Ref.

Post-socialist -11.287 ** (3.498)

Nordic 1.142 (3.639)

Conservative -2.869 (3.679)

Mediterranean -2.795 (3.424)

Index of Globalization23

0.377 * (0.156)

Voice and Accountability24

0.774 (2.090)

Constant 93.361 *** (12.162)

***p<001, **p<.01, *p<.05, ~p<0.1, two-tailed tests.

In sum, initial variance component analysis provides some doubt as to whether multilevel

analysis is necessary to explain variation in perceptions of the effectiveness of legal or

nonviolent political activity among young people. At least the variance component analysis

combined with the results of the multilevel model, suggest that a different set of contextual

variables is necessary to explain residual individual-level variation. However, the model

(corroborated by the unreported results of the fixed effects model with only individual level

indicators), suggests that use of the internet for information and political knowledge is

positively and substantively associated with increased perceptions of effectiveness of legal

or nonviolent activity. Several other individual-level variables exhibited a statistically

significant, but substantively small effect on these perceptions, providing little support for

strictly economic modernisation or political culture theories of political activity and

governance. There is some support that increased wealth and participation in employment

or education are also associated with increased perceptions of effectiveness, but these

effects are also substantively small. This result, especially taken in the context of

inconclusive effects of the contextual level variables, provides little support for hypotheses

that social, political, or economic dislocation drive young people to be more cynical about

constructive political participation.

21

Harmonized variable computed based on Q28_1, Q28_2, and Q28_3; 0=don’t know/incorrect on all three Q28 items (lowest knowledge), 3=answers to all Q28 items correct (highest knowledge) 22

Q29, ranging from 0=highest trust (“most people can be trusted”) to10=lowest trust (“you can’t be too

careful”). Note that higher values actually correspond to less trust. 23

Index of globalization (Dreher, 2006; Dreher et al., 2008); 0-100, higher values=higher levels of economic, social, and political globalization 24

World Bank Worldwide Governance Indicator; -2.5 to 2.5, higher scores=better governance.

MYPLACE August 2014


8.4 Summary

This chapter presents analysis conducted on the Q19 items of the MYPLACE survey, which

asks respondents to rate their perceptions of the effectiveness of a series of political

activities in influencing politics in their countries.





The analysis proceeded in three main stages. First, we conducted a factor analysis to reduce

the number of indicators to underlying latent variables. The factor analysis produced a two-

factor solution, with Q19 items tracking on to two separate scales: perceptions of the

effectiveness of legal or nonviolent political activity; and perceptions of the effectiveness of

illegal or violent political activity. We constructed both scales accordingly for use in future

analysis. Second, we conducted a descriptive analysis of all Q19 items and scales by survey

location. Individual items and scales did exhibit clustering by locations, although it is difficult

to identify patterns without statistical tests. In general, voting in elections received the

highest mean response for effectiveness of political action, followed by gaining publicity

through media exposure and being active in a political party. Participating in violent and

illegal protest activities are the least effective.In particular, the two Georgian locations

tended to rate all Q19 items as relatively effective while Russia and the Central European

countries (including Hungary, Croatia and Slovakia) tended to be relatively pessimistic about

the effectiveness of Q19 activities. Third, we conducted a multilevel regression analysis to

explore individual-level and group-level determinants of variation in the effectiveness of

legal or nonviolent activity scale. The results of the model suggest some support for the

hypotheses that the use of the internet for information and political knowledge were

associated with increased perceptions of effectiveness of legal or nonviolent activity.

However, the model provided little support, although did not conclusively falsify, major

hypotheses about youth political activity, including those derived from economic

modernisation theory, and the hypothesis that supranational political and economic change

may lead young people to be more pessimistic about the effectiveness of constructive

political participation. In terms of the goals of the MYPLACE project, this result seems to

indicate that accession to supranational political and economic institutions like the EU has

not had any significant effect, either positive or negative, on young people’s attitudes

toward the effectiveness of legal or nonviolent political activity. However, given the

concerns about the model elaborated above, further research is necessary to bring

additional evidence to bear on this hypothesis.

MYPLACE August 2014


References

Almond, G. A, and Verba, S. (1963) The Civic Culture: Political Attitudes and Democracy in

Five Nations. Newbury Park, Calif.: Sage Publications.

Boix, C. and Stokes, S. C. (2003) “Endogenous Democratization.” World Politics 55 (4): 517–

49. doi:10.1353/wp.2003.0019.

Dreher, A., Gaston, N. and Martens, P. (2008) Measuring Globalization – Gauging its

Consecquences. New York: Springer.

Duin, T. (1997) “The Decline of the SPD among Young People: A Study of Causes.” German

Politics 6 (3): 103–19. doi:10.1080/09644009708404491.

Edwards, K. (2009) “Disenfranchised Not ‘Deficient’: How the (Neoliberal) State

Disenfranchises Young People.” Australian Journal of Social Issues (Australian Council of

Social Service) 44 (1): 23–37.

Epstein, D. L., Bates, R., Goldstone, J., Kristensen, I. and O’Halloran, S. (2006) “Democratic

Transitions.” American Journal of Political Science 50 (3): 551–69. doi:10.1111/j.1540-

5907.2006.00201.x.

Forbrig, J. (2005) Revisiting Youth Political Participation: Challenges for Research and

Democratic Practice in Europe. Council of Europe.

Horvath, A.and Paolini, G. (2014). “Political Participation and EU Citizenship: Perceptions

and Behaviors of Young People.”

http://eacea.ec.europa.eu/youth/tools/documents/perception-behaviours.pdf.

Huntington, S.l P. (1968) Political Order in Changing Societies. New Haven, CT: Yale

University Press.

Inglehart, R. (1997) Modernization and Postmodernisation: Cultural, Economic, and Political

Change in 43 Societies.

Kääriäinen, J. andLehtonen, H. ( 2006) “The Variety of Social Capital in Welfare State

Regimes – a Comparative Study of 21 Countries.” European Societies 8 (1): 27–57.

doi:10.1080/14616690500491399.

Kaufmann, D., Kraay, A. and Mastruzzi, M. (2009) “Governance Matters VIII: Aggregate and

Individual Governance Indicators for 1996–2008”. World Bank Policy Research Paper No.

4978. http://ssrn.com/abstract=1424591

Lipset, S. M. (1959) “Some Social Requisites of Democracy: Economic Development and

Political Legitimacy.” American Political Science Review 53 (01): 69–105.

doi:10.2307/1951731.

MYPLACE August 2014


O’Donnell, G. A. (1973) Modernization and Bureaucratic-Authoritarianism: Studies in South

American Politics. Vol. 9. Institute of International Studies, University of California Berkeley.

http://www.getcited.org/pub/101476732.

Przeworski, A., Alvarez, M. E., Cheibub, J. A.and Limongi, F. (2000) Democracy and

Development: Political Institutions and Well-Being in the World, 1950-1990. 1st ed.

Cambridge University Press.

Putnam, R. D., Leonardi, R. and Nanetti.R. Y. (1993) Making Democracy Work: Civic

Traditions in Modern Italy. Princeton University Press.

Taft, J. K. and Gordon, H. R. (2013) “Youth Activists, Youth Councils, and Constrained

Democracy.” Education, Citizenship and Social Justice 8 (1): 87–100.

Winston, F. (2013) “Decisions to Make a Difference: The Role of Efficacy in Moderate

Student Activism.” Social Movement Studies 12 (4): 414–28.

doi:http://dx.doi.org/10.1080/14742837.2013.827569.

MYPLACE August 2014


Chapter 9: Citizenship

Literature

Citizenship Studies, as a separate multi-disciplinary field, was formed at the end of the 20th

century (Hoffman, 1999). At the beginning of the 21st century, there was an upsurge in the

research of citizenship (Stevenson, 2001). Traditional understandings of citizenship were put

to the test by the new political, economic and cultural conditions: globalisation and

mobility, neoliberalism and the neoconservative turn, which occurred in many countries

(Heater, 1999; Ellison, 1997). Nowadays, citizenship is conceptualised by a whole range of

disciplines: such as, sociology, political science, law, history, geography, pedagogy and

others. One of the classical definitions of citizenship is ‘a formal status of a national

affiliation of a person in state, implying a clear set of universal rights, implying equality of all

members of the state‘ (Marshall, 2006). Being a complex and multiple phenomenon,

citizenship is analysed through various components and manifestations (Ellison, 1997;

Crowley et al., 1997:1-3).

Ellison states three key directions in the analysis of citizenship: state-centred, pluralistic, and

post-structural (Ellison, 1997). ‘State-centred’ researchers understand citizenship in line

with Marshall’s definition – through the common rights and duties of a citizen, which he or

she gains being a citizen of a nation state. Additionally, two ways of understanding relations

between the state and citizens can be defined: liberal and republican (Oldfield, 1990; Miller,

1990).

Liberal theorists (Kymlicka, 1995; 2001) emphasise individual freedom and the right to live

one’s own life, as long as it does not interfere with the rights of other people (Mouffe,

1992:6). According to this understanding, cultural minorities are accepted in the liberal

society – they are allowed to preserve their own values and are not forced to share any

values that could be alien to them. State interference in individuals’ lives is justified mainly

as a way to maintain the rule of mutual respect - ‘respect the rights of others‘. Enhanced

integration of nation states and the need to regulate co-existence of multiple cultural

models, within a single state, have led to the emergence of multiculturalism - both in its

‘hard’ (Kymlicka, 1995) and ‘soft’ versions (Kukathas, 2009). Ideally, this approach has to

support cultural and ethnic diversity and maintain their conflict-free interaction. The

critique of this idea is concerned with the risk of equalising cultural differences and

reproduction of the dominant culture (Walzer, 2000). Zizek has argued that the neutrality

of multiculturalism is only the ’mask‘ of Eurocentrism (Interpassivity, 2005).

‘Republicanism‘ (Pettit and Skinner, 1997). According to this idea, freedom can exist only if

there is a suitable mode of law that protects citizens from violent interference and

MYPLACE August 2014


despotism. In the republican tradition, civil participation is an essential element of

citizenship. ‘True’ citizens should be interested in the common good and are willing to

contribute to it (Honohan, 2002: 5).

There is also a ‘communitarianist‘ position, whose ideas are to some extent similar to the

followers of ‘republicanism' (Walzer,2000). There is the need for a new form of sociality,

which would combine social duty and a desire for social good with personal autonomy.

Communitarians have accused liberals of excessive individualism and low social

responsibility, stating that they exaggerate the rights of citizens over the duties.

Followers of the ‘pluralist tradition‘ criticise the universal understanding of citizenship and

emphasise unequal access to rights and political participation (Young, 1989).Within the

frames of post-structuralism, the concept of citizenship is redefined as one of the aspects of

identity politics. Post-structuralists stress its dynamism, plurality and dependence on various

interpretations articulated in the context of democratic societies of late modernity (Mouffe,

2005; Seidman, 1999).

Based on these approaches of understanding citizenship, we can highlight several key

definitions:

1. Citizenship is a term that describes people’s belonging to a community and their

membership in it, as well as exclusion of the ‘other’ – ‘non-citizens’, ‘aliens’. Therefore,

citizenship is one of the most powerful ways of border-making in the contemporary world.

Brubaker is one of the first thinkers who pointed to this role of citizenship. In ‘Citizenship

and Nationhood in France and Germany’ he states that ‘citizenship itself by its very nature

has not only inclusive, but also exclusive status. Globally, citizenship is an extremely

powerful tool for social closeness. It protects prosperous and peaceful states, if there would

be no borders, from the vast majority of those who would want to escape from war, civil

strife, hunger, unemployment, environmental degradation, or decided to change their place

of residence in the hope of providing better opportunities for their children. The possibility

of obtaining citizenship is limited everywhere...’ (Brubaker, 1992; 2004). In this case, the

question of citizenship concerns not only formal membership in the society/state, but

politics of identity and democratic principles, also defined as ‘politics of recognition and

respect‘ (Lister, 2003:37; Hopkins, 2011).

2. Citizenship is a term that describes both passive membership and active participation. On

the one hand, one can be a citizen by just having a formal status without public participation

for the common good and the use of granted rights (e.g. the right to vote). Researchers in

recent years have pointed to the decline in interest to participation in politics (Putnam,

1999; Lister, 2002). On the other hand, we should assume that the status of a citizen, not

only refers to rights, but also the duties to be performed by a person. (See, for example,

MYPLACE August 2014


Oldfield, 1990). Being a citizen means to be active and to act for the common good. This

very understanding of participation and citizenship is also undergoing significant changes.

Today, people prefer not to act in the field of traditional politics, but rather in the fields of

culture, consumption, and the Internet (Stevenson, 2003; Delanty, 2000).

3.- Nowadays, citizenship describes legal political membership and activism as well as non-

political involvement in a community and non-political participation based on common

values. Additionally, citizenship is increasingly becoming associated not only with rights and

obligations, but also with certain pleasures. Riley et al.(2010) have even introduced the

concept of a ‘citizen of fun‘, which is used to describe new forms of youth communities

based on the values of hedonism. By analysing leisure activities, the authors conceptualise

personal, local and informal aspects of citizenship. In particular, they state that leisure is

increasingly becoming an important area for the construction of new identities (Haste,

2004).

4. Citizenship can be also regarded as a characteristic of a particular individual, indicating

one’s individual status and personal interests. On the other hand, it appears to be a trait of

collectivity (of one’s belonging to the community) (Gaventa, 2009).

Another discussion on citizenship considers ‘territorial rootedness‘ and citizens' mobility

(Urry, 2012). The link between citizenship and the nation state is not always relevant today.

Thus, according to some authors, citizenship is not only a matter of official belonging to the

state and living on its territory, it is more about subjective meanings, feelings and

performative strategies of being a citizen, in different communities at different levels: local,

national and global (Gaventa, 2009). One of the outcomes of globalisation today is the

emergence of a ‘global citizen‘ or a ‘’citizen of the world‘ who has overcome one’s own

rootedness in a certain territory, nation, ideology, family, tradition; he or she is mobile

among the countries without being emotionally bounded to places or people.

These avenues of understanding of citizenship in the westernern academic context show

that this category can be used to describe contradictory phenomena. At the same time,

their meanings are closely connected. In fact, the term ‘citizen‘ was conceptualised for the

strict distribution of rights, assets and duties in society. Yet in any contemporary

approaches to citizenship, deconstructing, rethinking and criticising this link between

citizenship and the nation state, do not overcome this concept’s scope, which, a-priori,

implies the existence of at least some identity connected to the state (Sassen, 2002). This

again brings us back to the construction of ‘citizen’ within the nation state. The question

emerges: ‘Who can be considered a citizen?’ In many works there is a dichotomy between

‘ethnic‘ and ‘civic‘ nationalism. Kohn can be regarded as the founder of this approach (Kohn,

1944.), who considered Eastern Europe through ‘ethnic nationalism‘ and Western Europe

through ‘civic nationalism‘. Keating defines civic nationalism, as opposed to ethnic

MYPLACE August 2014


nationalism, as a collective decision rooted in the individual agreement, rather than

ascriptive identity. It is based on common values and institutions, as well as on the patterns

of social interaction. Anyone can join the nation regardless of their place of birth or ethnic

origins (Keating, 1996:5-6).

According to this idea, ‘ethnic‘ nationalism is driven by common origin and is more closed,

isolating, and involuntary. In contrast, ‘civic‘ nationalism is driven by political participation

and leads to openness, inclusion, and voluntarism (Ignatieff, 1993). The critique of this

approach (Yak, 2006; Brubaker, 1998; Kuzio, 2002; Kymlicka, 2001) stresses the impossibility

of distinguishing ‘pure’ types of this dichotomy, the ambiguity of the terms ‘ethnic‘ and

‘civic’ and the closeness and non-voluntarism of both cases, as well as the need to consider

cultural aspects of citizenship. Besides this, some authors insist that the constructs of civic

and ethnic nationalism are mutually exclusive (Miller, 2000), while others state that

citizenship is a complex phenomenon based on a mixture of ‘ethnic‘,‘civic‘ and ‘cultural‘

nationalisms (Brubaker, 1998). However, it is important to mention that both the followers

and critics of ‘ethnic‘ and ‘civic‘ dichotomy refer only to political and state discourses, while

no mention is made of the views of the country’s inhabitants on the question of who should

be considered a citizen. Reeskens and Hooghe (2010) focus their critique on this problem

and show how the dichotomy between the ’ethnic‘ and ’civic‘ is understood by the public.

The authors use Shulman’s (2002) approach to define the indicators of civic identity.

Table 9 Operationalisation of civic, cultural and ethnic identity (Shulman, 2002: 559)

(Content of national identity - Key indicators)

Civic Cultural Ethnic

Live on the territory Believe in dominant religion Ancestry, descent

Have legal citizenship status Speak national language Belong to the dominant

ethnic/racial group

Express will to join political community Share national traditions

Adhere to basic state ideology

Adhere to political institutions and rights

It can be seen that knowledge of the language in this dichotomy (‘ethnic vs civic’) becomes

an intermediate value because, on the one hand, the language is something that can be

learnt, regardless of origin. On the other hand, according to Brubaker, “... the linguistic

nationalism is only a private expression of ethnic nationalism. When the term ‘ethnic‘ is

understood broadly as ethno-cultural or just cultural in general, then the understanding of

the nation as a linguistic community, the demand for autonomy or independence on behalf

of such a community, limiting access to citizenship on the basis of knowledge of the

language and encouraging or requiring language study, publishing , radio broadcasting,

management or advertising in this language must be viewed as major, even paradigmatic

manifestation of ethnic nationalism“ (Brubaker, 2004).

MYPLACE August 2014


In recent academic discussion, there has been a weakening of the ‘nationalism’ concept and

an emergence of the idea of the post-national state and trans-national citizenship. This

approach stresses the disconnection between rights and national belonging, citizenship and

national territories (such as, denationalisation, deterritorialisation) as well as the rupture

between formal and emotional belonging (Sassen, 2002). As Habermas argued, nationalism,

as a form of cultural identity, no longer reflects contemporary reality, ’so the overcoming of

fascism creates a particular historical perspective, from which the post-national identity is

formed based on the principles of law and democracy‘ (Habermas, 1990: 154).

In connection with this, in citizenship studies particular attention is paid to the question of

European identity. It has been expected that citizens who identify themselves as ‘European’

will show mutual solidarity; they expect to be prepared to share resources together and

thus agree to pay for less wealthy fellow citizens (Duschesne et al., 2013). However, critics

argue that the European identity is an artificial and vague concept that does not really exist.

Many researchers suggest that, more often, EU citizens identify themselves with their

national state rather than with the European Union (Bauböck et al., 2009). ‘Promoters of

the political identity claim that a common cultural background in Europe is too weak to

encourage Europeans’ awareness of belonging to this community‘(Pichler, 2012). It also

shows the ambiguity of the perception of free movement in the integrated European space

by the residents of the EU, according to the migration policy. Some researchers argue that

there are no existing social mechanisms to protect the rights of migrants outside of the

nation state in order to implement the concept of post-nationalism (Malakhov, 2014).

Globalisation leads to the emergence of a phenomenon, which researchers have called the

devaluation of sovereignty. In a global market, nation states are lacking the resources for

building their own economy. The emergence of supra-national political and economic

structures leads to de-territorialisation and changing the nation state’s function as the

provider of peace, comfort and well-being. In this case there are two approaches to the

vision of the nation state. In the first case, welfare is not connected with care for the nation,

but with the concern for the economic interests of the state, which entails the building of

public policy in the context of global economic and political trends. In the second approach,

the protection of ‘national interests‘ refers to the confrontation with global trends; the

government’s policy is aimed at the protection of domestic economy and the security of

citizens (Lakić, 2011; Malakhov, 2005).

Taking into consideration a complex understanding of citizenship, which includes political

participation, civil activism, and interaction of different communities within one territory, in

our analysis, we concentrate on the value assets of citizenship and young people’s

understanding of the relationship between the state and its citizens, as well as the

relationship between the state in which they live and other countries.

MYPLACE August 2014


9.1 National identity (Q22)

We asked our respondents the following question:

Q22: Are you a citizen of your country?

Besides this, we included the following questions in our survey:

Q56: What is your country of birth?

Q57: What is your nationality?

Q58: What is your ethnic group?

Based on all four questions, we created a combined category that describes the identity of

the respondent in terms of: belonging to the majority (which includes the dominant

nationality in the country and dominant ethnic group/s in the country) and to the minority.

MYPLACE August 2014


Q22: National identity


In the MYPLACE survey, respondents were asked the question: Are you a citizen of this country? It was a dummy variable with ‘yes/no’ categories. The locations in this graph are given in decreasing order of the

number of respondents who answered that they are citizens of this

country.

Graph 9.1a: Respondents’ citizenship (by location)

Narrative

16,905 respondents - from all survey locations - answered this

question. 93% of them considered themselves to be citizens of their

country and 7% said that they are not citizens of their country.

According to the results of the analysis, it turned out that only in 5

localities is the number of non-citizens greater than 10%: Narva

(Estonia), Coventry (UK), Vic (Spain), Forstate & Jaunbuve (Latvia) and

Barreiro (Portugal). As can be seen on Graph 9.1a, an interesting

cluster (which includes Slovakia, Hungary, Georgia, Russia, Croatia,

Finland) shows the number of citizens close to 100%.

MYPLACE August 2014


Q22, Q56, Q57 & Q58: Majority/minority status


This variable was created during the process of harmonising the data

from all the participating countries. The decision was taken by the

MYPLACE project leader. Based on 4 questions; Are you a citizen of

your country? What is your country of birth? What is your

nationality? What is your ethnic group? A unified category was

created, which describes a respondent’s identity in terms of:

belonging to the majority (which includes the dominant nationality in

the country and dominant ethnic group/s in the country) and to the

minority.

Graph 9.1b: Respondent’s majority/minority status (by location)

Narrative

The current index that describes a respondent’s identity covered

16,679 respondents. Based on the criteria we used, overall, 80% of

the respondents represent majority groups in their location, and 20%

represent minority groups.

We can highlight several countries with equal or almost equal number

of respondents that can be assigned to the minority using the index.

These are Georgia, Finland, Croatia and western Germany. In other

countries the difference between localities is vast. Especially, it can be

noticed in Estonia, Slovakia, Latvia and the United Kingdom.

MYPLACE August 2014


9.2 Feeling proud of current citizenship (Q24)

The existing traditional construct of citizenship as the formal belonging of a person to the

state, has excluded the person themselves from the process of citizenship ‘uptake‘. Such an

approach neglects the meaning of subjectivity of a citizen, paying attention only to accepted

roles. In this sense the importance of perception of oneself as citizen of a country becomes,

in some ways, a construct through the prism of which identity citizenship can be formed.

Respondents were asked to evaluate this using a 4 grade scale (‘very proud‘, ‘quite proud‘,

‘not very proud’, ‘not proud at all‘):‘How proud you are to be a citizen of your country?’

MYPLACE August 2014


Q24: Feeling proud of current citizenship


The graph shows, by location, levels of pride in being a citizen of a given. It is

ordered from highest to lowest levels of pride. Responses of ‘very proud’

and ‘quite proud’ were combined in order to calculate the level of pride in

citizenship.

Graph 9.2: How proud are you to have this citizenship? (by location)

Narrative

16,349 respondents answered this question from all survey locations. 39% of

them said that they are ‘very proud‘ that they are citizens of their country;

40% answered with ‘quite proud‘; 15% chose the option ‘not very proud‘;

6% said that they are ‘not proud at all‘ to be citizens of their country.

The highest percentage of those who are very proud of their citizenship is in

Georgia (in both locations this answer was given by more than 80% of

respondents). The lowest level of proudness of one’s own state citizenship is

in eastern Germany (in both localities only 12-13% said that they are ’very

proud’).

We can highlight two opposite groups of countries: 1. Where the majority of

youth are proud of their citizen status. 2. Where almost half of the

respondents are not proud to be citizens. The first cluster consists of

Georgia, Finland, Denmark and the United Kingdom. The second one – Spain

and eastern Germany. In many countries, in both localities the percentage of

youth who are proud to be citizens is close to equal. Nevertheless, in Latvia

and Estonia in different localities the amount of respondents in the

combined coefficient (‘not very proud’ and ‘not proud at all’) differs by 20%.

It is interesting that particularly in these countries (see the Graph 9.1b)

there is the highest difference between localities on the merged index of

belonging to majority or minority.

MYPLACE August 2014


9.3 Important things relating to country citizenship (Q25)

Citizenship in a modern understanding, is inseparable from a person and from their values

and beliefs on what it means to be a citizen of one’s own country. In our research,

respondents answered the question: ‘Some people say the following things are important

for being a citizen of country. Others say that they are not important. How important do you

think each of the following is?’

Using the five-grade scale (‘very important’, ‘important’, ‘neither important nor

unimportant’, ‘not very important’, ‘not important at all’) the young people who

participated in the research had to evaluate the importance of each of the following

notions:

Q25_1: To have been born in [COUNTRY]

Q25_2: To be able to speak [THE NATIONAL LANGUAGE]

Q25_3: To have at least one [COUNTRY] parent

Q25_4: To respect [COUNTRY] political institutions and laws

In line with the theoretical approach that was examined earlier, during our analysis we

combined Q25_1: To have been born in [COUNTRY] and Q25_3: To have at least one

[COUNTRY] parent in one unified scale as an indicator of ethnic nationalism demonstration.

Q25_4: To respect [COUNTRY] political institutions and laws is analysed as an indicator of

civic nationalism as opposed to ethnic.

Variable Q25_2: To be able to speak [THE NATIONAL LANGUAGE] [NOTE: if more than one

national language, ask the national languages] will be reviewed separately.

MYPLACE August 2014


Q25 (Q25_1 & Q25_3): Ethnic nationalism


In the MYPLACE survey, respondents were asked what was important for being a

citizen of [country]. Ethnic nationalism was coded as a summary variable from ’to

have been born in [country]’ and ’have at least one ethnic [country] parent)’.

Ordinal by ordinal Gamma correlation 0.699 suggest there is a strong relationship

between the two variables and theoretically they can be combined. Analysis of the

reliability of the scale revealed a high level of Cronbach alpha values (α = 0, 787) in

all countries except Estonia (0.62). It confirmed that the items form a coherent and

reliable scale.

To be able to compare this value with the value of civic nationalism, we constructed

a scale based on the average value. Our scale has a minimum value of 0 (the lowest

level of support for the concept of ethnic nationalism) and the maximum value of 4

(highest level of support for the concept of ethnic nationalism).

Graph 9.3a: Average values of respondents’ support of the ethnic

conception of citizenship (by survey location)

Narrative

The average value of the level of respondents’ support for the concept of ethnic

nationalism across the entire data base is 2.44 (number of respondents - 16,770,

standard deviation - 1.17).

The strongest support for the ethnic approach to the definition of citizenship was

observed in Vyborg (Russia) - 3.22. The lowest level of support for this concept was

expressed by inhabitants of Bremen (western Germany) - 1.46.

With regard to the localities where respondents are more positive about ethnically

oriented citizenship, both localities in Georgia and Greece can be highlighted.

We can also select a stable group, which includes eastern Germany, western

Germany, Denmark and Spain, where young people demonstrated the lowest level

of significance of ascriptive indicators of citizenship.

In some countries, we can also observe a significant difference between the two

localities in the average levels of support for ethnic nationalism: In Latvia

(Agenskalns and Forstate & Jaunbuve) the difference is 1.00; in Finland (Kuopio and

Lieksa and Nurmes) = 0.79; in Estonia (Narva and Tartu) = 0.74; and in Slovakia

(Trnava and Rimavska Sobota) = 0.73.

MYPLACE August 2014


Q25_4: Civic nationalism


In the MYPLACE survey, respondents were asked what was important for

being a citizen of [country]. Civic nationalism was coded according to

answers to the variable ‘To respect [COUNTRY] political institutions and

laws‘. The scale has a minimum value of 0 (the lowest level of support for

the concept of civic nationalism) and the maximum value of 4 (highest level

of support for the concept of civic nationalism).

Graph 9.3b: Average values of respondents’ support for the civic

concept of citizenship (by survey location)

Narrative

The average value of the level of respondents’ support for the concept of

civic nationalism across the whole data base is 3.17 (number of respondents

- 16,803 , standard deviation – 0.97).

The strongest support for the civic approach to the definition of citizenship

is showed in both localities of Denmark: Odense Center (3.67) Odense East

(3.55). The lowest level of support for this concept was expressed by the

residents of Forstate & Jaunbuve - 2.59 (Latvia).

Regarding the localities where we can observe the most positive attitudes

towards civic oriented citizenship, both localities in Denmark and western

Germany can be highlighted.

However, Croatia, Spain, Russia and Slovakia form the group in which young

people felt the lowest level of significance of indicators of civic citizenship.

In most of the countries that participated in the study, residents of different

localities expressed similar views in relation to civic nationalism. However, in

Latvia (Agenskalns and Forstate & Jaunbuve) the difference between the

average values of support of the civic concept is 0.65; in Estonia (Narva and

Tartu ) = 0.53.

MYPLACE August 2014


Q25_2: To be able to speak the national language


The scale has a minimum value of 0 (‘the ability to speak the national

language is not important at all’) and a maximum value of 4 (‘the

ability to speak the national language is very important’).

Graph 9.3c: Average value of respondents’ support for being able to

speak the national language (by survey location)

Narrative

The average value of the level of support by respondents for the

concept of civic nationalism across the entire data base is 3.23

(number of respondents - 16,875, standard deviation - 0.92).

The ability to speak the national language was the most important for

youth in Tartu (Estonia) = 3.66. The lowest level for this was expressed

by the respondents in Forstate & Jaunbuve ( Latvia) - 2.25 .

In general, Estonia has a very interesting position: in this table, Tartu

and Narva appear to be maximally opposed. If in all other countries

the average difference is not more than 0.25, the Baltic countries

revealed the greatest variability of responses in individual localities. In

Estonia (Tartu and Narva) the difference between the average values

equals 1.27; in Latvia (Agenskalns and Forstate & Jaunbuve) = 1.01.

We can select a cluster, which includes both localities in Spain and

Croatia, where the ability to speak the national language was least

important as a description of the features of a citizen of the country.

MYPLACE August 2014


9.4 Attitudes towards national sovereignty and externalities

(Q26)

Citizenship as a legal concept, in some ways naturally implies the existence of state

guarantees to protect citizens, their welfare and other legal privileges. Another issue is how

equitable and universal such dissemination of guarantees is for individuals that are

somehow included in the community-state scheme. From this perspective, the definition of

who exactly should be granted the privileges begins with alignment of boundaries between

‘us‘ and ‘them‘. And here we are not interested in power discourse as the rationale for state

sovereignty existence, territorial protection and protectionist policy. The Important focus

for our analysis is the attitude of young people to participating in such political aspects of

citizenship.

In the study, respondents were asked to rate on a five grade scale (‘strongly agree‘, ‘agree‘,

‘hard to say whether agree or disagree‘ , ‘disagree ‘, ‘strongly disagree‘) their attitude to the

following statements:

Q26_1: [COUNTRY] should limit the import of foreign products in order to protect its

national economy

Q26_2: The availability of foreign films, music, and books greatly contributes to

(COUNTRY’S) national culture

Q26_3: [COUNTRY] should follow its own interests, even if this leads to conflicts with

other nations

Q26_4: Foreigners should not be allowed to buy land in [COUNTRY]

Q26_5: Membership of the European Union greatly benefits this country

Q25_5 variable cannot be used for construction of a combined index as the question was

not answered by respondents from all countries (for youth in Georgia and Russia this

variable is not applicable for the analysis).

Using the factor analysis of the main components with the Varimax rotation, on the basis of

their own decisions across the data base, we have identified two factors explaining 67.5 %

of the variance. The first factor was assigned for variables (Q26_1, Q26_3, Q26_4), the

second - Q26_2. Suitability analysis showed the significance of Cronbach alpha (α = 0,608).

There is no certainty that this scale is stable, as the factor analysis for individual countries

demonstrates the integration of the different variables in the factors and satisfactory results

from the combined scale (Q26_1, Q26_3, Q26_4), Cronbach alpha ( greater than 0.6 ) can be

seen only in two countries (Estonia α = 0.600 ; Hungary α = 0,611).

For this reason, we did not produce indices and analyse each of the items separately.

MYPLACE August 2014


Q26_1: Country should limit the import of foreign products in order to protect its national economy


The graph is compiled by combining the ‘strongly agree’ and the

‘agree’ responses.

Graph 9.4a: Country should limit the import of foreign products in

order to protect its national economy (by location)

Narrative

16,394 respondents answered this question across the whole data

base: 16% of them said that they ‘strongly agree‘ with the policy of

protectionism; 29% said they ‘agree‘; 25% were ‘undecided’; 24%

chose the option ‘disagree‘; 7% said that they ‘strongly disagree‘ with

the policy of protectionism.

The highest percentage of those who agree that the state should

protect the national economy by limiting the import of goods was in

Greece (total percentage of those who strongly agreed and agreed) in

Argyroupouli - 79%, in New Philadelphia - 77%). The lowest overall

percentage of those who agreed was observed in Jena (eastern

Germany) - 12 %.

There are two opposing groups of countries: 1. Where a significant

majority of surveyed young people support (‘strongly agreed and

agreed’) a policy of protectionism; 2. Where a significant majority to a

varying degree do not agree with the policy of protectionism. The first

cluster includes both localities in Greece, Croatia, Slovakia and

Portugal. The second cluster - eastern Germany, western Germany

and Denmark.

MYPLACE August 2014


Q26_2: The availability of foreign films, music, and books greatly contributes to national culture


The graph is produced by combining the ‘strongly agree’ and the ‘agree’

responses.

Graph 9.4b: The availability of foreign films, music, and books

greatly contributes to national culture (by location)

Narrative

16,517 respondents answered this question across the whole data base:

17% of them reported that they ’strongly agree‘ with the positive effect of

the availability of foreign cultural ‘messages‘ on the national culture; 41%

said they ‘agree‘; 25% were ‘undecided’; 14% chose the answer ‘disagree‘;

and 3% ‘strongly disagree.’

The highest percentage of those who agree that the availability of foreign

films, music, and books contributes to the development of national culture,

was in Coventry (UK). The cumulative percentage (‘strongly agreed and

agreed’) was 81%. The lowest overall percentage of those who agreed was

identified in Vyborg (Russia) - 37%.

There are two opposing groups of countries: 1. Where a significant majority

of surveyed young people support (‘strongly agreed and agreed’) the

opinion about the positive impact of foreign cultural models on the

development of national culture; 2. Where a significant majority to a varying

degree ‘do not agree’ with this statement. The first cluster includes both

localities in Portugal, eastern and western Germany and the UK. The second

cluster - Russia and Georgia.

MYPLACE August 2014


Q26_3: [Country] should follow its own interests, even if this leads to conflicts with other nations



responses.

Graph 9.4c: [Country] should follow its own interests, even if this

leads to conflicts with other nations (by location)

Narrative


14% of them said that they ‘strongly agree‘ with the importance of

protecting the sovereignty of the country; 31% said they ‘agree‘; 26% were

‘undecided’; 23% chose the answer ‘disagree‘; and 6% ‘strongly disagree‘

that the state should be guided by its own interests, even if it leads to

conflicts with other countries.

The highest percentage of those who agree that the government should in

any case pursue its own interests was in New Philadelphia (Greece). The

cumulative percentage (‘strongly agreed and agreed’) was 80%. The lowest

total percentage of those who ‘agreed’ was found in Narva (Estonia) - 19%.


of surveyed young people consider the protection of the sovereign interests

of the country an important aspect (‘strongly agreed and agreed’); 2. Where

a significant majority to varying degrees ‘do not agree’ with this statement.

The first cluster includes both localities in Georgia, Greece, Hungary and

Portugal. A second cluster - both localities in Spain, eastern and western

Germany.

MYPLACE August 2014


Q26_4: Foreigners should not be allowed to buy land in [country]


The graph is produced by combining the ‘strongly agree’ and the


Graph 9.4d: Foreigners should not be allowed to buy land in

[country] (by location)

Narrative

16,396 respondents answered this question across the whole data

base: 11% said that they ‘strongly agree‘ that foreigners should not

have the right to purchase land; 18% said they ‘agree‘; 23% were

‘undecided’; 32% chose the answer ‘disagree‘; and 17 % said that they

‘strongly disagree‘ with the restriction of the sale of land to

foreigners.

The highest percentage of those who agree with this postulate

(‘strongly agreed and agreed’) was in Hungary (the total percentage in

Ozd - 64%, in Sopron - 58%). The lowest total percentage of those

who ‘agreed’ was observed in Jena (eastern Germany) - 1%.

We can distinguish two opposite clusters: 1. Where a significant

majority of surveyed young people supported restrictions for

foreigners. This includes both localities in Hungary, Slovakia and

Georgia; 2. Where a significant majority does ‘not agree’ with this.

This includes localities in eastern Germany, western Germany,

Denmark and Spain.

MYPLACE August 2014





responses.

The analysis excluded two countries: Russia and Georgia. Youth in Croatia

evaluated their country's future membership in the European Union.

Graph 9.4e: Membership of the European Union greatly benefits this

country (by location)

Narrative


11% said that they ‘strongly agree‘ with the fact that membership in the

European Union greatly benefits their country; 38% said they ‘agree’; 32%

were ‘undecided’; 14% chose the answer ‘disagree‘; and 5% ‘strongly

disagree.’

The highest percentage of respondents that ‘agreed’ that membership in the

European Union was recorded in Bremen (western Germany). The total

percentage (‘strongly agreed and agreed’) was 72%. The lowest total

percentage of those who ‘agreed’ was noted in two localities: Ozd (Hungary)

- 27%; Rimavska Sobota (Slovakia) - 27%.


of the young respondents positively evaluate EU membership (‘strongly

agreed and agreed’); 2. Where a significant majority to varying degrees ‘do

not agree’ with the fact that membership of the European Union greatly

benefits this country. The first cluster includes both localities in eastern

Germany, western Germany and Spain. The second cluster - localities in

Greece, Croatia and Hungary.

MYPLACE August 2014



Citizenship is a complex phenomenon, its interpretations and meanings are varied in

different societies and in different periods, ranging from the legal norms to political

philosophy and sociological concepts. Global transformations, the formation of

transnational and multicultural streams, and increasing migration invite researchers to

rethink the notions of citizenship and its practical implementation.

The issue of ‘de-territorialisation’ (‘de-nationalisation’) and the creation of supra-national

identities and communities, as we have already examined, in theoretical literature is one of

the key issues surrounding the modern concept of citizenship. Social researchers attempt to

define indicators, which allow us to evaluate and approach citizenship in terms of

isolation/inclusivity, openness/closedness, and rigid sovereignty/global cooperation. For

further analysis, we have chosen several variables, which strongly reflect these tendencies:

- Support for ’ethnic nationalism‘ (Graph 9.3.a)

- Attitude to the country joining the European Union (Graph 9.4.e)

- Support for the sovereignty of the country (Graph 9.4.c)

What follows is an analysis of the 3 models based on Multi-level regression analysis.

To understand the basic social meaning that young people give to the concept of

‘citizenship‘ in all three models, we consider individual understandings of citizenship from

several angles:

Through the status characteristics: gender (Q54), age (recoded variable – age1),

income (Q82), education (recoded variable - Q61R), nationality (harmonised

variable – Q57R), citizenship (harmonised variable – Q22R), social class (parents’

social class - harmonised variable, including parents’ education and profession

etc.).

Through the experience of personal identity: experience of feeling oneself a

citizen (Q24).

Through the construction of the ‘true‘ citizen (Q25).

Through the construction of the ‘right‘ country where one wants to live (Q26).

The parameters within the offered differentiation define the predictors of the individual

level in the three models.

We chose the following contextual predictors for model-building:

Indicators that characterise the economic status of the country:

o Per capita GDP (in US dollars)

Indicators that characterise the labour force of the country (unemployment rate,

population size in locations)

Net immigration rate

MYPLACE August 2014


Indicators of the political regime, which indicate the political/historical context of

the current situation in the country.

In the further analysis, we consider ordinal variables measured by a 5 and higher points

system factiously ordinal and continuous.

9.5.1 Model А – Ethnic ‘Nationalism’

Our dependent variable is ethnic ‘nationalism‘ (Graph 9.3.a), which, according to many

researchers, strengthens anti-migrant and xenophobic attitudes. In this case, the concept of

citizenship contains several meanings: formal status, belonging, exclusion and inclusion,

local and global, individual rights and collective solidarity, and images of the ’other‘. These

meanings can be contradictory sometimes, but this variance illustrates the complexity of

citizenship in contemporary society. As one of the hypotheses, we analysed interaction

among the three indicators of citizenship that we defined (Brubaker, 1998).

Independent variables25

The following is a short description of social factors, which determine young people’s

support of ‘ethnic nationalism’. Along with traditional socio-economic indicators like age,

gender, education, profession, social class – we included the country of birth, citizenship

and ethnic groups in our analysis. Research on citizenship gains much importance also in the

context of research on young people as a particular age group. Traditionally, citizenship has

been associated with attaining a certain age to have rights and duties including those in the

family and labour force, as well as a new autonomous and responsible status (Thomson et

al., 2004). However, in recent years this process has become problematic due to a high level

of youth unemployment and new post-material values and lifestyles (Bauman, 2002). In this

context, young people are more likely to stay living with their parents longer, and not hurry

to start their own family and have their own responsibilities, redefining the category of

‘adulthood‘ as well as the concept of ‘citizen‘. Our understanding of citizenship is closely

connected to the adult status and marked by life events, such as leaving home, getting a job,

starting a family, acquiring legal rights and responsibilities. While building the model, we

assumed that moving into adulthood is associated with an increased focus towards civic

nationalism.

In the latter part of the 1900s, the problem of conceptualising national identity was

characterised by the shift towards the definition of national identity in post-socialist

democratic transition (Central and Eastern Europe, Eurasia). Some researchers have

described the national identity of this region as ‘ethnic’ (Brubaker; Schopflin, 1995). For

25

A list of independent variables at the individual level for all models is included in Appendix 1.

MYPLACE August 2014


example, Brubaker concludes that in these countries, nationalist movements contributed to

the creation of the state and the state itself was created in the interest of a particular

ethno-political group. In line with this logic, we can assume that in the new democracies of

Central and Eastern Europe there is little chance of sharing the patterns of civic identity. In

the context of the removal of the Communist ideology, by other opposing ideas and values,

ethnicity has remained the only factor that is present in the structure of the national

identity in the region (Schopflin, 1995).

As we have already noted above, ‘citizenship‘ refers both to the inclusion and exclusion of

people. It has become one of the most powerful ways of constructing boundaries in the

contemporary world. Nowadays, in any society, one can find marginalised groups, whose

rights to ownership and recognition can be questioned. In our models, we described these

groups through migrants and ethnic minorities. In particular, as the independent variables,

we considered the presence of friends of a different ethnicity in their close environment, a

composite index of the attitude towards minorities, and the feeling of threat caused by

respondents’ belonging to ethnic or religious minorities. Based on our findings, as a

hypothesis, we assumed that the positive experience of interethnic interaction minimises

the support for the concept of ethnic nationalism.

The variable ‘negative attitude to minorities‘ used in the building of the model is composite

from 7 items, which were recoded from the point of view of negative attitude:

Roma, Gypsies and travellers make a positive contribution to society (Q40_1)

The police should be stricter with Roma / Gypsies / travellers (Q40_2)

Jewish people talk too much about what happened to them during the Holocaust

(Q40_3)

Jewish people make an important contribution to society (Q40_4)

Muslims make a positive contribution to society (Q40_5)

It is right to be suspicious of Muslims (Q40_6)

Migrants greatly contribute to national cultural diversity (Q41_3).

Factor analysis composes all the 7 items in one factor explaining 41.3% of the variance.

Reliability analysis shows a high value of Cronbach alpha (α = 0, 756).

According to many researchers, xenophobia is often based on the lack of adequate

knowledge and understanding of politics. Therefore, besides the interest in politics based on

some theoretical notions (Pettigrew 1971: Fetzer 2000), we included in our model the

presence of wide extended social networks, which include ethnic minorities.

As a hypothesis, while choosing individual-level predictors in our model, we assumed that,

in accordance with the widely shared views, ethnic nationalism is also strongly associated

with political and social trust. There is a theoretical approach that asserts that a strong

MYPLACE August 2014


national (ethnic) identity is not necessarily translated as support for, or credibility of,

national political institutions, and does not necessarily lead to a sense of solidarity with

fellow-citizens. In other words, nationalists are ready to strongly support their fellow

citizens (also materially) (Duschesne et al, 2013).

According to Snyder, ethnic nationalism tends to dominate when in the context of collapse

of state institutions, when they are no longer able to fulfill the basic needs of citizens .

Besides this, a critical attitude towards the state power is considered an integral part of

citizenship, for which the public interest prevails all other interests.

Results

Having selected independent variables for our analysis, we built a matrix of pair correlations

to determine the statistical relations between the dependent variable and potential

predictors, as well as to control multi-collinearity. We excluded variables with a low (even

statistically significant) correlation coefficient (or Cramer coefficient for nominal variables).

We began our analysis by building the ‘empty‘ model (only random intercept) to get the

individual and intercept variance components. By counting the intra-class correlation

coefficient (Heck et al., 2010), we found that 26 % (ICC = 0.259) of the variance in ethnic

nationalism is explained by the differences between the localities where the research was

conducted. Further, 74 % of the variance is related to individual differences between

respondents. The led us to use the method of hierarchical regression.

Table 9.5.1.1 Variance components of regression models explaining ethnic nationalism

Empty model

(only random

intercept)

1st model

(model with

predictors of

individual level)

2nd model (with

predictors of both

individual and

contextual levels)

Ethnic nationalism

σ (individual

level)

1.017258 0.699242 0.699246

σ (country level) 0.357275 0.114173 0.057324

-loglikelihood 48036.387 26864.964 26844.879

At the next stage of our analysis, we included the individual-level predictors described

above, which are significantly correlated with the dependent variable. After calculating via

the formula of intra-class correlation coefficient (ICC = 0.14), we conclude that the addition

of these independent variables reduced the residual, individual level variance by 12%. To

compare the advantages of the models we use the method of maximum likelihood

estimation. The difference between the loglikelihood of the two models (empty model and

the one with individual predictors) is distributed according to chi square. According to the

MYPLACE August 2014


table of critical values, the decrease in the residual variance is statistically significant ( df =

28, loglikelihood = 21171.423). At the same time, we see that most of the cross-country

dispersion (68%) is explained by individual differences between respondents: R2 = 0.6804.

However, part of the variations associated with the data on locality insert, remains

unexplained.

At the next stage of the analysis, we included the contextual predictors described above,

which led to the decrease in residual, individual level variance by another 5.5% (ICC = 0,65),

which is a statistically significant result ( df = 5 , loglikelihood = 20.085 ), so this caused a

decrease in cross-country dispersion by 49%. The results are shown below.

Table 9.5.1.2. Determinants of ethnic nationalism

Parameter Estimate

Standard

deviation Df. t Sig.

Constant term 1.445394 .489653 34.739 2.952 .006


[No interest in politics Q2_1=.00] .002245 .033384 10766.967 .067 .946

[Weak interest in politics Q2_1=1.00] .032818 .027741 10761.676 1.183 .237

[Moderate interest in politics Q2_1=2.00] -.003422 .027421 10756.054 -.125 .901

[Strong interest in politics Q2_1=3.00] 0a 0 . . .

[No friends of other ethnicities Q2_1=.00] .178708 .022087 10773.945 8.091 .000

[1 friend of a different ethnicity Q33_1=1.00] .190583 .026699 10776.518 7.138 .000

[2 friends of a different ethnicity Q33_1=2.00] .054856 .027009 10764.940 2.031 .042

[3 and more friends of a different ethnicity

Q33_1=3.00]

0a 0 . . .

[Not proud to be a citizen Q24=.00] -.518853 .040823 10782.936 -12.710 .000

[not very proud to be a citizen Q24=1.00] -.343801 .027873 10782.780 -12.334 .000

[quite proud to be a citizen Q24=2.00] -.210370 .019981 10778.184 -10.529 .000

[very proud to be a citizen Q24=3.00] 0a 0 . . .

[very difficult to live with such income

Q82=.00]

-.008409 .039304 10779.113 -.214 .831

[difficult to live with such income Q82=1.00] .003969 .026444 10779.982 .150 .881

Fine to live with such income Q82=2.00] .016294 .020509 10775.319 .794 .427

[Comfortable to live with such income

Q82=3.00]

0a 0 . . .

Negative attitude towards minorities (Q40 &

Q41_3)

.209965 .017251 10782.031 12.171 .000

The country should tighten border control to

control migration (Q41_1)

.085436 .008736 10780.887 9.779 .000

Migrants without citizenship should have same

access to social security as citizens (Q41_2)

-.077911 .008814 10772.444 -8.839 .000

Strong leader of the country not restricted by

the Parliament (Q43_1)

.048741 .007426 10782.610 6.564 .000

Satisfied with life (Q27) -.004163 .004376 10775.067 -.951 .341

Religiousness (Q37) .020370 .002985 10782.985 6.823 .000

MYPLACE August 2014


Recognition of importance to remember the

past of the country (Q53)

-.009728 .004259 10780.031 -2.284 .022

Citizen/non-citizen (1,0) (Q22) .268874 .040778 10782.221 6.594 .000

Dominant ethnicity/non-dominant ethnicity

(1,0) (Q58)

.180983 .029864 10731.285 6.060 .000

Protectionism – the country should restrict

import of foreign goods (Q26_1)

.071808 .008179 10777.869 8.779 .000

Age -.010396 .003026 10780.620 -3.435 .001

The level of social trust Q29 -.005450 .003686 10782.097 -1.479 .139

[never felt threat due to belonging to a

different ethnicity Q48_2=.00]

.086410 .077840 10755.914 1.110 .267

[sometimes felt threat due to belonging to a

different ethnicity Q48_2=1.00]

-.011971 .082266 10754.348 -.146 .884

[regularly felt threat due to belonging to a

different ethnicity Q48 _2=2.00]

0a 0 . . .

It is important to speak the national language

to be a citizen (Q25_2)

.392936 .010565 10781.499 37.192 .000

Civic nationalism (Q25_4) .050339 .009584 10772.284 5.252 .000


Population size in a locality (in thousands of

people)

.000001 .000001 29.505 .635 .530

GDP per capita

(in US dollars)

-.000042 .000012 29.708 -3.513 .001

Unemployment rate (%) -.010098 .011689 29.385 -.864 .395

net_immigration_rate .038397 .032463 29.572 1.183 .246

Post-socialist countries (1,0) -.219011 .167963 29.595 -1.304 .202

We find no significant relationship between the perception of family welfare and support

for ethnic nationalism; on the individual level, understanding of one’s own poverty and

financial difficulties in the family, was neither associated with attitudes toward the ‘other‘,

nor with the subjective satisfaction with life in general. However, the context of the state

financial situation significantly differentiates the understanding of the importance of

ascriptive characteristics, which define a citizen. According to the theory, the increase in the

country’s welfare reduces the probability of ethnic nationalism. In general, among all the

contextual variables included in our model, only the country's GDP per capita proved to be a

significant predictor. Contrary to the theoretical notions, according to which Kohn stated

the division between eastern and western nationalism (Kohn, 1944), our model

demonstrates the absence of such a relationship in accordance with the criticism of this

approach (Remizov, 2011). Neither unemployment rates, migration, nor population size in

the locality, significantly affect the support for ethnic nationalism.

We now turn to the analysis of the socio -demographic characteristics of respondents. Even

at the stage of building a matrix of pair correlations, we took away many indicators due to

their low correlation with the dependent variable (e.g. education, gender, employment,

social class calculated as a combined variable through parents’ education and employment).

MYPLACE August 2014


Among all the socio-demographic indicators included in our model, age has the smallest

effect on the support for an ethnic understanding of citizenship (the degree of impact is

based on Student statistics (t), as Estimates are not standardised and defined by the unit

variable). Nevertheless, growing older reduces the role of the place of birth in the definition

of a citizen, as we predicted in our hypothesis.

A dominant position in the country, in terms of having citizenship and belonging to the

dominant ethnicity, increases the probability of ethnic nationalism. This is supported by a

theoretical hypothesis of a higher importance of national identity for people belonging to

the dominant ethnic group (Lewin-Epstein, Levanon, 2005). In our study, only 13% of

respondents from all the countries indicated different ethnicity, and among them, 26%

reported that at least sometimes they felt threatened due to belonging to the minority. This

may explain the fact that this independent variable does not differentiate the attitude

towards ethnic nationalism. In contrast, positive experience of cooperation, such as broad

ethnic networks, reduces the support for an ethnic concept of citizenship.

Some researchers recognise a discursive link between citizenship and religion (Omel’chenko,

2012). In our model, respondents confirmed the influence of religiousness on support for

the ethnic concept of citizenship: more religious people are more likely to consider

ascriptive characteristics as important for citizenship.

With regard to the influence of the cognitive component, this turned out to be statistically

insignificant. Recognition of the importance of cultural and historical heritage has a weak,

but decreasing effect on the dependent variable.

In analysing the relationship between ideas of the ‘right‘ structure of the world and support

for ethnic concept of citizenship, we admit that our model does not allow us to understand

the causal relationship, but only a pattern of interaction between the variables.

Value assets that determine the importance of formal citizenship are in a significant linear

relationship with our dependent variable. The lower the importance of citizenship, the

lower the importance of ethnic characteristics of citizenship. This finding is supported by the

research on Russian youth conducted by the research centre ‘Region‘(Russia) and the Centre

for Youth Studies (Russia). Their research shows that the nation state is recognised as an

‘imagined community’ (Anderson, 2001) that constitutes individual identity, according to

which the value of being a citizen of the nation, leads to the protection of its borders. A

logical and theoretically relevant finding is the observed relationship between negative

attitudes towards minorities and the choice of ‘ethnic citizenship‘. An increase by 1 unit of

negative attitude leads to an increase by 0.203 units of the significance of ascriptive

characteristics in the definition of the ‘right‘ citizen (both variables are measured by a 5-

point scale ranging from 0 to 4).

MYPLACE August 2014


Young people’s support for the closed nation state, focused on the minimisation of

migration, support for the national economy against international global economic streams,

segregation of migrants, and their distancing from the system of social security - increases

the probability of ethnic nationalism. As we have noted above, the post-socialist status of

the country is not significant in this case. However, the support for totalitarian autocratic

principles of the political system enforces ethnic nationalism.

Let us refer to our hypothesis on the interaction between all the three defined indicators of

citizenship. Coefficients of determination support this hypothesis. Thus, the ability to speak

the national language has the strongest effect on the dependent variable. This indirectly

supports the theory of Reeskens and Hooghe (2010), who in their study noted that the

language of nationalism is not present. At the same time, in accordance with Brubaker’s

argument, the growth of civic nationalism leads to the increase of ethnic nationalism.

9.5.2 Model В – Membership of the European Union greatly

benefits this country

The emergence of supra-national social institutions and the hegemony of rights of a human

and not a citizen, which are actively promoted by contemporary political and philosophical

thinkers, leads one to attempt to rethink and outline the borders of emerging transnational

identities.

From this point of view, it is interesting to analyse the respondents' assessment of the

significance of their countries joining the EU. Dependent variable - ‘the importance of

membership in the EU‘ (Graph 9.4.e), in the questionnaire it was asked as ‘Membership of

the European Union greatly benefits this country’ (Q26_5). The participants had to agree or

disagree with this statement. Respondents from 13 countries answered this question

(excluding Russia and Georgia).


In building the model, we used traditional indicators of SES and socio-demographic status.

Some studies state that people in managerial position, as well social activists, support the

idea of European integration and justify all the problems it may provoke. However, most of

the ‘ordinary’ citizens do not share this view. ’Less privileged people do not like the EU

because they fear a loss of national sovereignty and the breakup of the welfare state. They

do have these fears, but they are related more generally to globalisation, not specifically - or

in some cases, at all - to European integration. Regarding integration, what characterises

working class people is their sheer lack of interest in Europe, the evident sense that they

MYPLACE August 2014


simply do not pay attention to it, that they overlook it. European issues are just not salient‘

(Citizens' Reactions…, 2013).

Therefore, we assumed that socio-economic status is one of the important characteristics,

which is closely related to the support for EU membership. It partially supports the

theoretical assumption that higher levels of well-being and the high presence of the middle

class reduces the significance of nationalist ideas.

The idea of a united Europe is based on the assumption that all the citizens of European

countries consider themselves primarily as ‘Europeans’. Some scholars consider that

European identity is a phenomenon that was brought up in opposition to national feelings.

However, most people believe that European identity and national identity, in contrast, are

compatible and are positively correlated: ‘It is clear that national identity is not competing

with feeling oneself as European: on the contrary, national identity is for most citizens a

template for European identification. People (albeit with some notable exceptions) feel

European because they feel German, French, Italian, Spanish, Polish, Czech, etc.‘ (Citizens'

Reactions…, 2013). It comes from a study of national or European identity in the structure of

the dominant.

There are two research positions describing the attitude of citizens of EU member states to

membership in the EU. One position states that dissatisfaction with EU membership is due

to a lack of confidence in the supra-national institutions, in terms of their ability to solve the

problems of the country (Baranowsky, 2006). According to another point of view, on the

contrary, dissatisfaction is not driven by the economic and social processes of EU

membership, but by a low degree of trust in their national political institutions, as opposed

to the credibility of European institutions (Alexandrov, 2008).

Level of political trust was calculated as a unified variable, including trust in:

The head of government (Q7_3)

Parliament (Q7_8)

Political parties (Q7_13).

Reliability analysis shows a high value of Cronbach alpha (α = 0.847).

We also included in the model such indicators as respondent’s interest in events happening

in their city, country, and the world. The variable is composite from 5 items - How interested

would you say you are in issues concerning:

the neighbourhood you currently live in (Q1_1)

the city you currently live in (Q1_2)

the country you live in (Q1_3)

countries neighbouring the country you live in (Q1_4)

Europe (Q1_5)

Factor analysis composes all the 5 items in one factor explaining 56.7% of the variance.

Reliability analysis shows a high value of Cronbach alpha (α = 0.807).

MYPLACE August 2014


Results

After selecting independent variables for our analysis, we built a matrix of pair correlations

to determine the statistical relations between the dependent variable and potential

predictors, as well as to control multi-collinearity. We excluded the variables with a low

(even statistically significant) correlation coefficient (or Cramer’s coefficient for nominal

variables) from the analysis.

We began our analysis with building an ‘empty‘ model (only random intercept) to get the

individual and intercept variance components. The intra-class correlation coefficient was

0.077, which means that 92.3% of the variance is explained by the individual differences of

the respondents and only 7.7% is determined by the differences in the localities where the

research was conducted. These results suggest that contextual variables cannot explain the

dependent variable. To test this hypothesis, we built several models with different

contextual predictors (including GDP, population in the locality , net immigration rate, time

of membership in the European Union, Gini coefficient, Human Development Index), but the

analysis showed that there was neither a significant improvement in the model nor a

reduction of the residual, individual level variance. Therefore, in the analysis we focus on

the building of the first model, which includes only individual predictors.

Table 9.5.2.1. Variance components of regression models explaining the importance of

membership in the EU

Empty model (only random

effect from the locality input)

(1st model) the model with

individual-level predictors

Attitude towards

the country’s

membership in

the EU

σ (individual level) 0.968973 0.859515

σ (country level) 0.081788 0.044315

-loglikelihood 39475.551 29454.635

At the second stage of the analysis, we included the individual-level predictors in the model,

which have significant correlation with the dependent variable. Having calculated using the

intra-class correlation coefficient (ICC = 0.049) formula, we can note that the addition of

these independent variables reduced the residual, individual level variance by 3%. To

compare the models we used the method of maximum likelihood estimation. The difference

between the loglikelihood between the two models (empty model and individual predictors)

is distributed by chi square. According to the table of critical values, the reduction of

residual variance is statistically significant (df = 23, loglikelihood = 10020.915). In this

case, we see that intercept variance decreased by 45%.

MYPLACE August 2014


The model we produced is below:

Table 9.5.2.2. Determinants of importance of membership in the EU

Parameter Estimate

Standard


Constant term 1.084262 .103251 840.145 10.501 .000


[Very difficult to live with such income Q82=.00] -.100239 .041985 10844.743 -2.387 .017

[Difficult to live with such income [Q82=1.00] -.094289 .029557 10821.673 -3.190 .001

[Fine to live with such income Q82=2.00] -.055513 .022905 10898.462 -2.424 .015

[Comfortable to live with such income Q82=3.00] 0a 0 . . .

[Other occupation Q61=.00] .074822 .041706 10925.825 1.794 .073

[Students Q61=1.00] -.012052 .035219 10931.148 -.342 .732

[Non-workers Q61=2.00] .052782 .021839 10932.291 2.417 .016

[Workers Q61=3.00] 0a 0 . . .

[Rights are not at all upheld Q47=.00] -.273453 .047166 10933.492 -5.798 .000

[Right are almost never upheld Q47=1.00] -.158716 .031310 10928.198 -5.069 .000

[Rights are considerably upheld Q47=2.00] -.070142 .026686 10932.467 -2.628 .009

[Rights are completely upheld Q47=3.00] 0a 0 . . .

The country should tighten border control to control

migration (Q41_1)

-.016257 .009161 10854.183 -1.775 .076

Migrants with no citizenship should have equal

access to social security like citizens (Q41_2)

.079291 .009438 10931.174 8.401 .000

Political trust (president. parliament. parties)

(Q7_pol)

.009138 .001986 10706.393 4.601 .000

Satisfaction with life (Q27) .020514 .004829 10931.259 4.248 .000

Interest in events (Q1) .000863 .003343 10927.809 .258 .796

The country has a democratic multiparty political

system (Q43_2)

.067516 .011065 10927.319 6.102 .000

The country has an oppositional movement with a

freedom to express opinion (Q43_4)

.007750 .011490 10933.109 .674 .500

Civic nationalism (Q25_4) .064991 .010782 10931.699 6.028 .000

Ethnic nationalism (Q25_1 & Q25_3) .031368 .010053 10481.019 3.120 .002

Ability to speak the national language (Q25_2) .015430 .012021 10901.154 1.284 .199

Trust in the European Commission (Q7_7) .088931 .005219 10931.489 17.041 .000

[Low social class3=.00] -.068719 .025654 10933.005 -2.679 .007

[lower middle social Class3=1.00] -.056589 .033595 10929.537 -1.684 .092

[upper middle social Class3=2.00] -.075310 .033368 10917.883 -2.257 .024

[High social Class3=3.00] 0a 0 . . .

We considered the hypothesis that dissatisfaction with EU membership is due to the lack of

confidence in pan-European or supra-national political institutions or, conversely, life

satisfaction is not associated with membership of EU, but with the negative evaluation of

national political structures.

MYPLACE August 2014


However, both the hypotheses were only partly confirmed in our model. The strongest

effect (according to Student statistics (t)) on the recognition of the importance of

membership was confidence in the European Commission as an institution. The higher the

level of trust in the European Commission, the more positive the young people’s assessment

of the country’s membership in the EU. At the same time, the growth of confidence in the

national political institutions (government, parliament, political parties) causes an increase

in the support for EU membership.

According to researchers of European identity, our model shows that the increase in life

satisfaction leads to an increase in support for joining the EU. Similar logic is found in the

subjective understanding of one’s own economic well-being: the less satisfying it is, the

more negative attitude towards joining the EU the person has.

Additionally, the social capital of the respondent, employment, social class of parents, does

not have a significant effect on the dependent variable. It is also worth noting that the idea

of violation of human rights in the nation state only reduces the positive attitude towards

the EU. This may occur due to the loss of trust in social institutions that guarantee human

rights in general: if a country cannot provide an individual with human rights it means that

supra-national institutions have even less chance of providing them.

There is a theoretical hypothesis that the ethno-cultural factor in the EU functions as a

deconstruction of its unity, strengthening the dichotomy between a United Europe vs

Europe of nations (Berendeev, 2012; Popova, 2002). This idea about the relationship

between national identity and attitudes towards the countries joining the European Union

was not confirmed. If we consider ‘ethnic‘ and ‘civic‘ components that imply the national

construct of citizenship, the importance of civic and ethnic nationalism increases the

positive attitude towards EU membership. Also, the support for the democratic political

system and the absence of anti-immigrant ideas also increases the importance of joining the

European Union. This may create a basis for the further construction of transnational

identity.

9.5.3 Model C – The support for sovereignty

We have already noted the duality of citizenship, which describes characteristics of the

individual and community (for example, a nation state). In the latter characteristic of

citizenship the emphasis is on the construction of the ‘right‘ community and the world

around it. In the Russian case, such a duality of meanings is grasped by the concepts of

‘citizenship‘ and ‘nationality‘, where the first determines individual rights, and the second

the community’s characteristic.

MYPLACE August 2014


The collective element of citizenship includes relations with other countries, that is, the

distribution of power within the state and outside. From this perspective, in our study the

protection of the sovereign rights of the state has crucial meaning. However, the

sovereignty and the freedom of each state in international relations are limited by the

freedom of other states. In our study, we approach the concept of sovereignty by ignoring

possible conflicts with other countries.

Our dependent variable ‘the support for sovereignty’ (Graph 9.4.c) – was measured by the

five-point scale of agreement/disagreement with the following statement: [COUNTRY]

should follow its own interests, even if this leads to conflicts with other nations (Q26_3).

Independent variable

Quite often this dichotomy becomes the framework of thinking about the concepts of an

opened/closed state and democratic/authoritarian forms of rule. To understand these

relations, we included individual predictors, such as attitude towards minorities, support for

protectionism, and approval of a strong leader in the country not constrained by the

Parliament. We also considered belonging to the dominant nationality as a controlling

factor.

Researchers define citizenship as a part of identity politics, which acquires complex and

situational meanings. A nation ceases to be the only ‘imagined community’ and one’s own

identity can be based on any other category (religion, economic views, knowledge, ethnicity,

ideology). Two forms of interaction with the state are constructed:

- People feel national and care for it because they depend on what public policies provide

for them in terms of protection and empowerment. As Norbert Elias (2000) explained,

feeling as ‘we‘ is the manifestation of an identity that first relates to people’s units of

survival, to the communities, concreted and imagined, that give them the resources to go

on. Nations remain the template for political identification because they provide their

citizens with resources that secure their social rights (Elias, 2000). There is a shift from the

support for the nation state to more individual identities (Malakhov, 2005).

- One’s own identity is created through the attitude to the state, to its power structures, and

through the reproduction of ideas about the ‘ideal‘ country, in which one wants to live. In

our research we asked the respondents not only about their formal citizenship, but also how

proud they are to be citizens of their countries. This independent variable is used to assess

the impact of nation-oriented personal values on the nation-oriented state policy.

Researchers have already stressed the existing inequality in access to the rights among

different population groups (Lister, 2003). In this case, the question of citizenship concerns

MYPLACE August 2014


not only formal membership of a society/state, but the recognition of a particular group in

society and, in general, the ‘politics of recognition and respect ‘. Therefore, it is interesting

to analyse how sharing the position of defense of the national sovereignty is connected with

the respect for multiplicity of identities and equal rights as a condition of citizenship. As the

independent variables in the building of our models, we took the respondents’ opinions on

whether migrants without citizenship should have the same access to social security as

citizens, and whether the country should tighten border control to prevent migration.

In the end, we also included welfare state types, described by Kääriäinen and Lehtonen

(2006), in our model as contextual factors. As we have mentioned in Model A, in post-

socialist transition countries (Central and Eastern Europe, Eurasia), according to some

researchers, the creation of the state was followed by nationalist movements, and the state

itself was created in the interests of certain ethno-political groups, which increase the will to

preserve the nation-oriented policy (Schopflin, 1995).

Results

After selecting independent variables for our analysis, we built a matrix of pair correlations

to analyse the statistical relations between our dependent variable and potential predictors,

as well as to control multi-collinearity. After the preliminary analysis, we excluded the

variables that had a low (even statistically significant) correlation coefficient (or Cramer’s

coefficient for nominal variables).

After building an ‘empty‘ model (only random intercept) and calculating the intra-class

correlation coefficient, we found that 14.3% (ICC = 0.143) of the variance in the attitude

towards the need to protect the country’s sovereignty can be explained by the differences

in the localities where the research was conducted. 85.7% of the variance refers to the

individual differences among respondents.

Table 9.5.3.1 Variance components of regression models explaining the support for the

defense of the country’s sovereignty

Empty model

(only random

effect from

the locality

input)

(1st model)

model with

individual-level

predictors

(2nd model)

Model with

individual and

contextual

predictors (data

about the

countries)

The country should

be sovereign

σ (individual

level)

1.092738 0.958771 0.958823

σ (country level) 0.183178 0.081680 0.021057

-loglikelihood 48344.967 30481.924 30444.586

MYPLACE August 2014


Next, we included the individual-level predictors, which have significant correlation with the

dependent variable. We can state (ICC = 0.785) that adding independent variables led to the

decrease in residual, individual level variance by 6.4 %. To compare the advantages of the

models we used the maximum likelihood estimation method. The difference between the

loglikelihood of the two models is distributed by chi square. According to the table of critical

values, the decrease of the residual variance is statistically significant ( df = 18,

loglikelihood = 17863.043). At the same time, we see that almost a half of the inter-

country variance (55%) is explained by individual differences among respondents: R2 =

0.554. However, some variety associated with the input of locality remains unexplained.

At the next stage of the analysis, we included contextual predictors, which led to the

reduction in residual, individual level variance by other 5.8% (ICC = 0.214). Therefore, there

is only 2.1% of unexplained variance. The reduction of the variance is statistically significant

at the confidence level of 0.005 ( df = 7, loglikelihood = 37.338). This caused a decrease in

the cross-country variance by another 74%.

Below is the model that we produced:

Table 9.5.3.2. Determinants of the need to protect the country’s sovereignty

Parameter Estimate

Standard


Constant term 2.702644 .296272 35.005 9.122 .000


[not proud to be a citizen Q24=.00] -.254487 .046216 10741.793 -5.506 .000

[not very proud to be a citizen Q24=1.00] -.113755 .031503 10816.744 -3.611 .000

[quite proud to be a citizen Q24=2.00] -.066987 .022942 10865.958 -2.920 .004

[very proud to be a citizen Q24=3.00] 0a 0 . . .

[very difficult to live with such income Q82=.00] .038886 .045511 10451.088 .854 .393

[difficult to live with such income Q82=1.00] -.034389 .030749 10562.659 -1.118 .263

[possible to live with such income Q82=2.00] .002286 .023801 10862.545 .096 .923

[comfortable to live with such income Q82=3.00] 0a 0 . . .

Negative attitude towards minorities (Q40 & Q41_3) .189235 .019844 10760.378 9.536 .000

The country should tighten border control to prevent

migration (Q41_1)

.110376 .010059 10834.337 10.973 .000

Migrants should have the same access to social

security as citizens (Q41_2)

-.001553 .010231 10864.938 -.152 .879

The country has a strong leader not constrained by

the Parliament (Q43_1)

.060959 .008577 10805.951 7.107 .000

Satisfied with life (Q27) -.010146 .005115 10862.604 -1.984 .047

Religiousness (Q37) -.009143 .003425 10801.360 -2.670 .008

Protectionism – a country should limit imported

goods (Q26_1)

.210449 .009479 10835.330 22.202 .000

MYPLACE August 2014


Political trust (to the President. Parliament. political

parties) (Q7_pol)

-.004094 .001670 10820.893 -2.451 .014

Nationality (Q57) .099163 .041362 6337.962 2.397 .017

[different occupation Q61=.00] .037604 .043496 10865.331 .865 .387

[students Q61=1.00] .032423 .036866 10864.828 .879 .379

[non-workers Q61=2.00] -.041210 .022974 10855.100 -1.794 .073

[workers Q61=3.00] 0a 0 . . .


Population size in the locality (in thousands of

people)

-.000001 .000001 28.385 -1.322 .197

GDP per capital (in US dollars) -.000048 .000007 30.324 -6.834 .000

Unemployment rate (%) .004837 .009396 27.900 .515 .611

[Post-socialist WST=1.00] -.458448 .174641 28.282 -2.625 .014

[Nordic WST=2.00] .460491 .137035 28.472 3.360 .002

[Conservative WST=3.00] .658365 .163215 28.313 4.034 .000

[Mediterranean WST=4.00] -.244113 .191702 27.825 -1.273 .213

[Liberal WST=5.00] 0a 0 . . .

In the analysis of national sovereignty, it is particularly interesting to consider it in the

context that most of the countries participating in the research are included in the European

Union as a supranational structure. In political research the question of the transfer of part

of the national sovereignty to the European Union is widely debated. Contrary to the idea of

inevitable devaluation of national sovereignty in the context of globalisation, according to

some authors, the member states of the European Union keep most of their sovereign rights

concerning the most important issues of political and economic life. ‘As the director of the

Institute for European Studies in Brussels Paul Magnette states, in the four main tasks of the

state (border protection, maintenance of national identity, definition of the political

structure of society, and market regulation) only the latter is affected by the supranational

regulation from Brussels‘(Bordachov, 2007).Young people's ideas about the need for the

further protection of the sovereignty of the nation state can be regarded as a follow-up

discussion about the possibility of a supra-national citizenship and supranational

institutions.

The first thing we notice in our model is an important connection between xenophobic

sentiments and the dependent variable. The stronger negative attitudes towards minorities

and anti-migration views are, the more important it is to protect national interests despite

possible conflicts with other countries.

Protection of sovereignty, as independence from other states, is connected with economic

independence. Therefore, protectionism has the strongest effect on the dependent variable.

According to the respondents, protection of the national economy through the restrictions

of imported goods is closely connected with the defense of the country’s sovereignty:

agreement with the first idea entails agreement with the second.

MYPLACE August 2014


With regard to the connection between political trust and national sovereignty, a high level

of trust in political institutions reduces the support for national sovereignty.

Our analysis confirmed the hypothesis that the commitment to a more totalitarian regime

largely determines the attitude to the country’s independence. At the same time, the

analysis of contextual variables, in which we included welfare state types, shows the

following trend: for post-socialist countries the importance of national sovereignty is much

weaker than for Nordic and Conservative countries.

Satisfaction with life reduces the importance of sovereignty of the state. However, from the

point of material well-being, macro-economic indicators of the country’s welfare become

more important rather than individual financial difficulties.

Concerning the assumption that ethno-cultural factors are a differentiating phenomenon

based on national identity (Berendeev, 2012), we also find that belonging to the dominant

nationality increases the respondents’ agreement with the dominance of the state’s

interests. As we expected, the importance of the country’s sovereignty is associated with a

higher level of national pride: a higher level of pride of being a citizen determines a higher

level of significance of the dependent variable. In eastern Germany and Spain, the survey

showed young people’s lowest level of pride for their citizenship (the same for both

localities), and in these countries the support for sovereignty is much lower than in other

countries.

9.6 Summary Global transformations of contemporary society, the formation of transnational and

multicultural flows and migration processes are changing the traditional understanding of

citizenship as the formal belonging of an individual to the nation state. In this regard, in our

study we focused on various forms of contemporary citizenship, including certain values and

experience of self-determination. Taking into account the whole variety of interpretations

and meanings, we focused on the value assets of national identity, and young people’s

views on global and local state relations in cultural, economic and political dimensions.

Formation of one’s own civic views and position, is inextricably connected with the values of

young people and perception of oneself as a citizen of one’s country. Those who are most

proud of their citizenship status26 are the young people from locations in Georgia, Finland,

26 Our analysis is sensitive to nuances across our research locations where nationality, ethnicity and citizenship




nationality and ethnic status. Overall 80% of respondents can be attributed to the dominant majority26

of the

MYPLACE August 2014


Denmark and the UK with levels of pride around 90%. Approximately 60% of the young

people from Vic and Sant Cugat (both locations are in Catalonia) state that they are not

proud of their citizenship. Other locations where levels of pride in citizenship were less than

70% are: the Narva area (Estonia), Bremen (western Germany) Jena and Rostock (both in

eastern Germany), and Forstate & Jaunbuve (Latvia).

In general, in both western and eastern Germany, Denmark and Spain, young people placed

low importance on ascriptive indicators of citizenship. The strongest meaning of ethnic

citizenship was found in Vyborg (Russia). The ethnic role of citizenship is also strongly

supported in all the localities in Georgia and Greece. However, there is a strong difference

between localities in the following countries: Latvia (Agenskalns and Forstate and

Jaunbuve), Finland (Kuopio and Lieksa/Nurmes), Estonia (Narva and Tartu) and Slovenia

(Trnava and Rimavska Sobota). In these countries, indicators of support for ethnic

nationalism between the two different localities are significantly different.

In our study there are the two opposing groups of countries: one in which a significant

majority of young respondents fully, or partially, support the policy of protectionism

(Greece, Croatia , Slovakia and Portugal); and a second comprising those who do not agree

with it (eastern and western Germany, Denmark). For example, 77-79% of respondents in

the two localities in Greece support the idea that the state should protect its national

economy by limiting imported goods. In contrast, for example, in eastern Germany, the

proportion was only 12%. Support for the protection of sovereignty ranges from 19% in

Narva (Estonia) to 80% in New Philadelphia (Greece). The idea that the ‘country should

follow its own interests even if this leads to conflicts with other nations’, is supported by the

majority of respondents in locations in Georgia, Greece, Hungary and Portugal.A significant

majority do not agree with this position in Spain, eastern and western Germany.

The highest percentage of those who support restrictions for foreigners, for example, on

their right to purchase land, is in Hungary (from 58% to 64% of the respondents in both

localities). A similar view is supported in Georgia and Slovakia. In general, throughout the

whole dataset, this view is supported by 29% of young people. Thus, for example in eastern

Germany, in Jena, only 1% of people support this view. In eastern and western Germany,

Denmark and Spain, a significant majority did not agree with it.

Finally, an important factor in the understanding of citizenship is the evaluation of

membership of the EU. The overall results of the study demonstrated young people’s

country, i.e. not belonging to any minority citizen, national or ethnic group within the country though this

varies significantly, in Estonia for example, in the Narva area only 9% are classified as part of the majority

whereas in Tartu the figure is 96%.

MYPLACE August 2014


support for such integration (the total percentage of the full or partial agreement was 72%).

The highest rate was found in Bremen (westernern Germany), the lowest - in Ozd (Hungary)

and Rimavska Sobota (Slovakia) - 27%. A significant majority of young respondents positively

evaluate EU membership in eastern Germany, western Germany and Spain. However, in

Greece, Croatia and Hungary young people mostly disagree with it.

If we sum up these trends, we can identify a cluster of countries where most young people

share the idea of the nation state in its classical sense: Greece, Hungary, Slovakia, Portugal,

Georgia. There is also a group of countries where young people prefer other state

formations that are not rooted in the protection of "national interests”.

The availability of foreign films, music and books contributes to the development of national

culture received the greatest support in Coventry (UK). Again, there is a clear difference

between countries where a significant majority of young people see the positive impact of

foreign cultural models on the development of national culture (Portugal, eastern and

western Germany, UK) and countries, where young people do not support this view (Russia

and Georgia).

In our further analysis, using Multilevel Modeling, we analysed three models, which, in our

opinion, have the strongest differentiating impact on the approach to citizenship: support

for ‘ethnic nationalism‘; ’the importance of the country’s membership in the EU‘; and ‘the

importance of the protection of national sovereignty’.

The first model was based on the analysis of social factors that lead to young people’s

support for ethnic nationalism. It should first be noted that there is an intersection of

ethnic, civil and cultural indicators of citizenship. Thus, the ability to speak the national

language and its relevance for civic nationalism, is associated with ethnic nationalism.

Contextual characteristics of the position of the country and its well-being increase or

decrease the probability of ethnic nationalism. Of the contextual variables included in our

model, the country's GDP per capita has the highest impact on the importance of ethnic

citizenship: the higher the GDP, the higher the support for ethnic nationalism.

Being a citizen of the country and belonging to the dominant ethnic group increases the

probability of ethnic nationalism. The importance of the status of citizenship for a young

person is in a significant linear relationship with support for the concept of ethnic

nationalism. The lower it is, the lower the importance of ethnic characteristics of citizenship.

There is also a direct relationship between negative attitudes towards minorities and

preference for the ’ethnic concept‘ of citizenship. In contrast, the positive experience of

inter-ethnic interaction and wider inter-ethnic networks reduces the relevance of the ethnic

concept of citizenship.

MYPLACE August 2014


Young people’s support for a closed image of the nation state, strict control of migration,

protection of the national economy against global economic flows and segregation of

migrants increase the probability of ethnic nationalism. Contrary to our expectations, the

post-socialist status of the country was not significant. However, support for totalitarian or

autocratic state forms reinforces ethnic nationalism.

The socio-demographic characteristics of the respondents (education, gender, age,

employment, social class) have no significant effect on the adoption of the ethnic concept of

citizenship.

The second model was based on the assessment of the importance of EU membership.

Confidence in pan-European or supra-national political institutions has an interesting effect

on the evaluation of membership of the EU. Trust in the European Commission leads to

support for EU membership. However, the same trend is found in relation to the national

political institutions (government, parliament, political parties).

Violation of human rights in the nation state only reduces positive attitudes towards

membership of the EU. This may occur due to the general loss of confidence in social

institutions as providers of human rights. A positive attitude towards EU membership not

only increases social capital (employment, social class), but also a subjective view

onmaterial well-being and the overall level of satisfaction with life.

The importance of both civic and ethnic nationalism increases the positive attitude towards

EU membership. Again this confirms the importance of both ’ethnic‘ and ‘civic’ components

in the understanding of citizenship.

At the same time, support for a democratic political system and the lack of anti-migrant

sentiments also leads to the positive assessment of membership of the EU. This may create

the basis for the further construction of transnational identity.

The third model is based on the understanding of the prioritisation of national sovereignty.

Support for the idea of sovereignty is directly related to the xenophobic attitudes of our

respondents. Strong anti-migration sentiment leads to stronger support for the protection

of the ‘national interests‘, such as the need to be guided by public interest despite possible

conflicts with other countries (and this applies both to the political and economic

sovereignty). At the same time, belonging to the dominant ‘group’ and satisfaction with the

status of a citizen of the country increase respondents’ agreement with the dominance of

state interests. The level of trust in the national political institutions also has an interesting

effect; a low level of trust increases the will for political and economic sovereignty

MYPLACE August 2014


In contrast, the factor of the country's financial well-being is very significant; the more

financially successful the state, the less support for the idea of the national sovereignty.

Moreover, it is macro-economic indicators that are important here - individual poverty or

wealth do not play a key role. Similarly, an increasing level of satisfaction with life in general

reduces the importance of the protection of the sovereign rights of the state.

References

Anderson, B. (2001) Imagined Communities. Moscow: Canon Press - C, Kuchkovo Field,

Baranowsky, V.G. (2006) European Community in international relations.

Bauböck, R., Perchinig, B., Sievers, W. (2009) (eds). Citizenship Policies in the New Europe.

Amsterdam: IMISCOE / Amsterdam University Press.

Bauman, Z. (2002) The Individualized Society / Per. from English. ed VL Inozemtseva .

Moscow: Logos,

Berendeev, M. (2012) "European identity" today: the category of political practice or

discourse? Bulletin of the Baltic Federal University. Kant. 6. pp. 70-79

Bordachov, T. (2007) Evolution of the European Union and Russia-EU relations nowadays /

The Club of the world political economy. URL: http://www.wpec.ru/text/200708291310.htm

Brubaker, R. (1992) Citizenship and Nationhood in France and Germany. Cambridge, MA:

Harvard University Press.

Brubaker, R. (2004) Ethnicity without Groups. Harvard, Harvard University Press

Brubaker, R. (1998) Myths and Misconceptions in the Study of Nationalism. from The State

of the Nation. Ernest Gellner and the Theory of Nationalism. UK.

Duchesne, S., Frazer, E., Haegel, F. and Van Ingelgom, V. (eds) (2013) Citizens' Reactions to

European Integration Compared: Overlooking Europe. Basingstoke: Palgrave Macmillan

Crowley, H. et al. (1997) Citizenship: Pushing the Boundaries. Feminist Review. 57 (1)

Delanty, G. (2000) The Resurgence of the City in Europe?: The Space of European

Citizenship. In Isin. E. F. (ed.) (2000) Democracy, Citizenship and the Global City. London:

Routledge pp. 79-92

Elias, N. (2000) The Civilizing Process. Sociogenetic and Psychogenetic Investigations. Revised

edition. Oxford: Blackwell.

Ellison, N. (1997) Towards a new social politics: citizenship and reflexivity in late modernity.

Sociology. 31 (4): pp. 697-717

MYPLACE August 2014


Gaventa, J. (2009) Exploring citizenship, participation and accountability. IDS bulletin. 33 (2):

pp. 1-14

Habermas, J. des Neohistorismus, G.. Ein Gesprach mit JM Ferr i (1990) Kleine politischen

schriften. F / am-Mein.

Haste, H. (2004) Constructing the Citizen Political Psychology. 25 (3): pp. 413-439

Heater, D. (1999) What is Citizenship? Cambridge: Polity Press

Heck, R. H., Thomas S. L., Tabata L. N. (2010) Multilevel and longitudinal modeling with IBM

SPSS. New York: Routledge.

Hoffman, J. (2004) Citizenship Beyond the State. Sage

Honohan, I. (2002) Civic Republicanism. Routledge, London.

Hopkins, N., Blackwood L. (2011) Everyday citizenship: Identity and recognition / / Journal of

Community and Applied Social Psychology. 21 (3): pp. 215-227.

Ignatieff, M. (1993) Blood and Belonging: Journeys into the New Nationalism. N.Y: Farrar,

Straus and Giroux.

Interpassivity. Desire. Multiculturalism / Per. from English. Smirnova; ed. Mazin B. and G.

Rohonyan. St. Petersburg. : Aletheia, 2005

Kääriäinen, J. and Lehtonen, H. (2006) The Variety of Social Capital in Welfare Sate Regimes

‐ A Comparative Study of 21 Countries. European Societies. 8(1): pp. 27-57

Keating, M. (1996) Nations against the State: The New Politics of Nationalism in Quebec,

Catalonia and Scotland. New York, St. Martin's Press,

Kohn, H. (1994) The Idea of Nationalism. A Study in Its Origin and Background. London:

Macmillan.

Kukathas, C. (2009) Theoretical Foundations of Multiculturalism. URL:

http://inliberty.ru/library/study/327/

Kuzio, T. (2002) The myth of the civic state. A critical survey of Hans Kohn’s framework for

understanding nationalism. Ethnic and Racial Studies. 25(1): pp. 20–39

Kymlicka, W. (1995) Multicultural Citizenship: A Liberal Theory of Minority Rights. New York:

Oxford University Press

Kymlicka, W. (2001) Politics in the Vernacular: Nationalism, Multiculturalism, and

Citizenship. Oxford: Oxford University Press.

Lakić, N. (2011) Is globalization a challenge or a threat to nation-states as a dominant form

of polity? Western Balkans Security Observer 6 (21): pp. 6-17

Lewin-Epstein, N., Levanon, A. (2005) National identity and xenophobia in an ethnically

divided society. In UNESCO: National Identity and attitudes toward migrants. Findings from

the ISSP. International Journal in Multicultural Societies. 7 (2)

MYPLACE August 2014


Lister, R. A (2002) Politics of recognition and respect: Involving people with experience in

poverty in decision making that affects their lives. Social Policy and Society. 1: pp. 37 – 46

Lister, R. Citizenship (2003) Feminist Perspectives. 2 ed. Palgrave: Macmillan, 2003

Malakhov, V.S. (2005) Nationalism as a political ideology. Moscow: KDU.

Marshall, T. (2006) Citizenship and social class. The Welfare State Reader. Polity Press,

Bristol. pp. 30-39

Miller, D. (2000) Citizenship and National Identity. Cambridge: Polity Press.

Miller, D. (1990) Market, state, and community: Theoretical foundations of market socialism.


Mouffe, C. (ed.) (1992) Dimensions of Radical Democracy: Pluralism, Citizenship, Community.

Verso, London and New York.

Mouffe, C. (2005) The return of the political. Verso: London/New York

Oldfield, A. (1990) Citizenship and community: Civic republicanism and the modern world.

London: Routledge

Pettit, P. H. (1997) Republicanism: A Theory of Freedom and Government. Oxford University

Press, Oxford

Pichler, F. (2014) European Identity: Conceptual Approaches. URL: www.cinefogo. cuni.cz /

getfile.php? & id_file = 87

Popova, O.V. (2002) Political identity in the transformation of society. St. Petersburg, Univ

State University

Putnam, R. (1999) Bowling Alone. New York, Simon and Shuster

Reeskens, T., Hooghe, M. (2010) Beyond the civic-ethnic dichotomy. Investigating the

structure of citizenship concepts across 13 OECD countries. Nations and Nationalism. 16(4):

pp. 579-597.

Remizov, M. (2011) Russian civic nationalism against neo-feudalism. Issues of nationalism. 5.

Riley, S., Morey, Y. and Griffin, C. (2010) The 'Pleasure Citizen': Analysing Partying as a Form

of Social and Political Participation. Young. 18 (1): pp. 33-54

Sassen, S. (2002) Towards Post-National and Denationalized Citizenship in Isin E. and Turner

B. eds. (2002) Handbook of Citizenship Studies. Sage, pp. 277-292

Schopflin, G. (1995) Nationalism and Ethnicity in Europe, East and West. Nationalism and

Nationalities in the New Europe / CA Kupchan Ed. Ithaca, NY: Cornell University Press, 1995.

pp. 37 - 65

Seidman, G. (1999) Gendered Citizenship: South Africa's Democratic Transition and the

Construction of a Gendered State. Gender and Society (13) 3. pp. 287-307

MYPLACE August 2014


Shulman, S. (2002) Challenging the civic/ethnic and West/East dichotomies in the study of

nationalism. Comparative Political Studies. 35 (5)

Stevenson, N. (2001) (ed.) Culture and Citizenship. Sage

Stevenson, N. (2003) Cultural Citizenship in the Cultural Society: A Cosmopolitan Approach.

Citizenship Studies. 7 (3): pp. 331-348

Thomson, R., et al.(2004) Inventing Adulthoods: a Biographical Approach to Understanding

Youth Citizenship. The Sociological Review 52(2): pp. 218-239.

Turner, B. S. (1999) The Sociology of Citizenship. London: Sage

Urry, J. (2012) Sociology beyond societies: types of mobility for the XXI century. Ed.: Higher

School of Economics

Walzer, M. (2000) On Toleration. Moscow: Idea

Young, I. (1989) Polity and group difference: A critique of the ideal universal citizenship.

Ethics. 99: pp. 250-274

Malakhov, V. (2014) Kul’turnye razlichiia I politicheskie granitsy v epokhu global’nykh

migratsii /Vladimir Malakhov: Novoe literaturnoe obozrenie. Institut filosofii.

Kozlova, A. A. (2005) (ed.) O patriotizme I grazhdanstvennosti. SPB: EleksisPrint

Omel’chenko, E. L. (2012) Kak nauchit’ lyubit’ rodinu? Diskursivnye praktiki patrioticheskogo

vospitaniia molodezhi. In S chego nachinaetsya rodina: molodezh’ v labirintakh patriotizma/

Eds. Omel’chenko, E. and Pilkington, H. (2012) Ul’ianovsk: Ul’ianovskii gosudarstennyi

universitetpp. 263-310.

Yak, B. (2006) Mif grazhdanskoi natsii. Prognozis 2 (6)

MYPLACE August 2014


Chapter 10: Social networks, gender relationships and

sexuality

Early in the MYPLACE questionnaire, the section ‘Social networks, gender relationships and

sexuality‘, contains two questions aiming at a general assessment of the society the

respondents live in were asked. The first one focuses on trust and the second one on

people`s readiness to help others. The main part of this questionnaire section consists of

three questions on the frequency and way of contact (in person or via social networks,

media) to close friends as well as the diversity between the respondents and their friends.

Further on, some provoking statements were implemented in this section yielding at

enlightening the respondent`s attitudes towards women`s and homosexuals` rights and

positions in society.

Key Findings on social networks, gender relationships and

sexuality of national reports

There is no general archetype to classify the empirical results related to social networks,

gender relationship and sexuality across all the fourteen countries and 30 research locations

included in the MYPLACE project. The analysis is initially undertaken at the locational level

as the data for each location is representative. Subsequent analysis examines the extent to

which there are similarities or differences between the locations within each country. There

are two important caveats here: firstly there are four research locations in Germany, two in

eastern Germany and two in western Germany, for this reason we talk of “15 countries” to

differentiate the data. Secondly, in aggregating the data in this way we do not assert that it

is nationally representative. It cannot be as these reseach sites were chosen purposively

and the data is only representative of them and not of any national unit. Hence we are

examining the extent to which these locations, selected to contrast with each other within a

country, show similarities or difference. With this qualification we are able to assert that:

none of the research hypotheses, that better educated young people, coming from higher

social classes,

have higher trust in other people,

are embedded in a network of friends with a highly diverse pattern are more supportive towards gender equality and

accept homosexuality and

accept the right of women for abortion concerned, have been falsified by the empirical findings.

However if we focus only on

MYPLACE August 2014


a single topic out of the three sub themes (either social networks, or gender relationship or sexuality) across all countries or

one case (= country), but then comprising all three sub themes (social networks, gender relationship and sexuality),

general patterns become more visible.

Last but not least, there were only slight differences between the two sites of each of the 15

countries, if any. Mostly these differences can be explained by different socio-economic

conditions of the sites.

First coming to the key findings related to social networks:

Here we distinguish between the network of friends and the general trust in people (= social

trust).

In five out of 15 countries the social trust of young people is comparably low (Croatia,

eastern Germany, Greece, Latvia, Russia). In all of the 15 countries, the general pattern is

that better educated young people, quite often accompanied by membership to upper

social classes, tend to have higher trust in other people. The only exception is eastern

Germany, where better educated, people have a lower level of social trust. Religious young

people have lower trust to other people (e.g. in Georgia).

Concerning the interaction with friends, in all 15 countries we find a relatively high level of

networks of friends. What is interesting is the level of diversity of these “mate” networks (in

terms of socio-economic backgrounds, political views, gender). It is interesting to note, that

diverse political views do not construct a criteria for inclusion/exclusion as well as the

gender, whereas the level of education is influencing very much the level of diversity of the

friends´ networks. Furthermore, members of upper social classes in general hold a higher

diversity in friendship networks.

Concerning the intensity of interaction with friends, young males, unmarried, in education,

have a higher level of interaction with close friends (UK case). Furthermore, there is a

positive correlation between the usage of social media and a broad network of friends

(Russia, western Germany).Finally but not least: The older the young people are, the

broader is their network (Finland).

Concerning the sub theme, gender relationship, the supporters of equal opportunities for

men and women in the labour market and in politics have a vast majority in all 15 countires.

However there a some major constraints, regarding to

male/female views

ethnicity

religion.

MYPLACE August 2014


Differences between young men and young women: In some countries, young women tend

to support gender equality (Finland, Georgia, eastern Germany) whereas in Greece and

Russia men generally support patriarchal views. Western Germany is an exception insofar,

as young women prioritise men in the labour market, when jobs are scarce.

Different ethnicities and religions could play an important role for the question of equal

opportunities: For example in Denmark, young Muslims have a less equalitarian attitude

than Christians or non-believers. In Latvia young people from the Russian minority again

have more conservative views about equal opportunities for men and women.

The sub theme sexuality has to be, split between the attitudes of young people towards

homosexuality and towards the right of women for abortion. In six countries (Croatia,

Estonia, Greece, Hungary, Russia and Slovakia), the majority of respondents are against

homosexuality. Of course there are differences between the cases: Women are more

tolerant with regard to homosexuality than men (e.g. western and eastern Germany,

Finland, Greece), as well as high educated youths (Croatia, UK). In contrast young Muslims

reject homosexuality more than young Christians and non-believers (UK, Denmark, Estonia).

The right of women for abortion is under dispute. In the majority of the 15 cases this right is

accepted, but again with important constraints: In general more young women than young

men are accepting this right, especially in Croatia, eastern Germany and the UK. In Greece

the majority of males disapprove of this right. Furthermore there is an ethnic/religious

division for this question. In particular Muslims reject this right (UK, eastern Germany).

In Russia attitudes are divided: 40 percent are against abortion.

Literature

Social Trust

The issues of trust and information are not only related to direct social interactions, but can

also be applied to the analysis of bureaucracy and democracy. Social trust constitutes the

foundation for social cooperation and, by extension, for the integration and stability of

modern society (Nannestad, 2008; Rothstein and Eek, 2009). Classic authors such as de

Tocqueville, J.S. Mill, Durkheim, Simmel, Tönnies, and Kornhauser relate trust to

membership in voluntary organisations and stress the social benefits that result from this

trust and behaviour for the community.27

Similar assertions are stated in Karen S. Cook’s edited book ‘Trust in Society’ (2003) by

several authors where they stress the many functions trust performs in social and political

life in diverse social settings, from the local and personal to the national and institutional.

27

For a review of the classic literature on trust see Misztal, 1996.

MYPLACE August 2014


Hardin (2000) explores some of the conceptual complexities of trust relationships. He

differentiates between two kinds of trust: trust that relates to the confidence a specific

individual places in another individual (encapsulated interest) and, the second, generalised

trustworthiness. Hardin explores the ways in which the government can promote trust in

the latter sense. Referring to Hume he also puts forward that, institutions “play a role in

underwriting interpersonal trust” (ibid. 30).

However, exploring the empirical association between social and political trust, and its

relationship with membership of voluntary organisations, Newton (2001) argues that the

relationship between social and political trust and democracy is more complex and indirect

than appears to be the case at first.

Other studies confirm the strong relationship. Tao et al. (2014) find strong evidence that

political trust enhances social trust in China by referring to a generalised social trust

question, as well as three contextualised trust questions. They also shed light on the impact

of economic modernisation on social trust.

When applying these theories about social trust to young people, Kelly (2009) says that

young people are particularly affected by social exclusion resulting from disparities in

educational settings, economic disadvantages, and health disparities are more likely to

experience and feel low social trust. He stresses that social trust is an important factor for

civic and social engagement and is therefore a significant trait for adulthood. His article does

not only offer a dense overview of literature but also focuses on young migrants in the U.S.

and factors that contribute to their inclusion into society.

Social Networks

The analysis of social networks “is suggested as a tool for linking micro and macro levels of

sociological theory”, as Granovetter (1973:1360) suggests in his famous article ‘The strength

of weak ties’. The analysis of processes in interpersonal networks allows us to translate

small-scale interactions into large-scale patterns. To explore how and where influence,

information, mobility opportunity and community organisation is diffused in social

interactions, Granovetter suggests to not only focus on strong ties between individuals

(thus, interaction within small well defined groups) but also to consider the weak ties – thus,

the relations between groups. Weak ties allow for a wider social structure and for a wider

exchange of ideas, values and innovations because bridging ties gives access to resources

and opportunities that exist in one network to a member of another network. The structure

and group dynamics of a network can either facilitate or constrain the individual’s ability to

gather information and to increase access to different forms of capital, particularly

intellectual capital. Social ties can facilitate undesirable behaviours and imprison actors in

risky social situations. Woolcock (2001) elaborates on “social capital” and how one’s family,

MYPLACE August 2014


friends, and associates constitute an important asset, not only in situations of crisis, but also

for material gain and economic development.

Newman and Dale (2005) put forward that not all social networks are, created equally.

“Networks composed of a diversity of “bridging” links to a diverse web of resources and

“bonding” links that build trust strengthen a community’s ability to adapt to change, but

networks composed only of “bonding” links can impose constraining social norms and foster

group homophily, reducing resilience. Diversity fosters the resilience needed to adapt to

unexpected change, and can also enlarge the ability to proactively make collective decisions

that optimize future options” (ibid 1).

Similarly, McPherson et al. (2001) analyse causes of homophily and refer to geography,

family ties, organisational foci, informal roles, cognitive processes. They stress that

“homophily“ limits people's social worlds in a way that has powerful implications for the

information they receive, the attitudes they form, and the interactions they experience.

Homophily in race and ethnicity creates the strongest divides in people’s personal

environments, with age, religion, education, occupation, and gender following in roughly

that order” (ibid. 415).

According to these theories and approaches, social networking site users have probably

more friends and more close friends. Exploring people’s overall social networks and how use

of new media technologies is, related to trust, tolerance, social support, and community and

political engagement, Hampton et al. (2011) conclude that users of social networking

technologies are associated with having a larger overall network.

Wimmer and Lewis’ (2010) remarkable work on racial homogeneity in a group of college

students who used Facebook shows that “homophily based on other attributes—including

socioeconomic status, regional background, and shared cultural taste—may intersect with

racial homophily if there is significant overlap in category membership” (ibid. 586). They

further stress that also balancing mechanisms such as the tendency to reciprocate

friendships or to befriend the friends of friends, can amplify the homogeneity effects of

homophily.

Gender Relationships and Sexuality

Engagement and trust in democracy is, closely linked to equality, diversity and inclusion. In

this respect, young people’s perceptions of gender roles, relationships and sexuality are an

indicator of their reflexivity on diversity and their tolerance as such. Discourses of gender

and sexuality show how young people position themselves and how far they are, affected by

gender and sexuality related violence (Rivers and Duncan, 2013).

MYPLACE August 2014


Herek (2009) presents a framework analysis on sexual stigma and prejudice and shows how

they develop as cultural constructs in societies. O’Neil and Gidengil (2006) put forward in

their book that gender is a social capital and can considerably affect political engagement

and career.

The acceptance of gender non-conformity varies according to young people’s contact with

gay and lesbian persons, as Collier et al. (2010) found out in their research. Similar findings

are also, suggested by Heinze and Horn (2009) in their study about the relationship between

intergroup contact and adolescents’ attitudes regarding homosexuality and the treatment

of lesbian and gay (LG) peers.

10.1 Social trust (Q29) and Views on helpfulness of others

(Q30), combined

The respondents` assessment of people`s trustworthiness and their helpfulness is measured

on an 11-point-scale (from 0 to 10). Thereby, the highest value stands for the most sceptical

attitude (‘can`t be too careful’ and ‘most people look out for themselves’).

For the following analyses, we recoded the values in order to have the highest value

standing for the most positive attitude. We considered constructing a scale referring to

respondent`s assessment of the society they live in by adding both questions. Cronbach`s α

had a value of 0.69 for this scale and a correlation test revealed a correlation coefficient of

0.53. Thus, both questions were, added in a scale (“society trust scale”) and are presented

together in this sub section.

MYPLACE August 2014


Q29+Q30: General assessment of society


Response distribution of the constructed scale is nearly normally distributed,

however, slightly right skewed (more positive assessments of people`s

helpfulness/trustworthiness).

Graph 10.1: Mean views of people`s trustworthiness and helpfulness

Narrative

The overall mean value for the society trust scale is 7.72 (n=16,792,

sd=4.10). Highest mean score is, given for Kuopio (Finland, 11.48) and the

lowest mean score value is, found in Telavi (Georgia, 4.42).

While both locations of Finland, western Germany and Portugal belong to

the top third, both locations of Denmark, Latvia and the UK can be found in

the middle field and both field sites of Greece, Russia, Slovakia and Georgia

are in the bottom third.

There are some countries with pronounced differences between field sites.

Eastern German respondents revealed a more positive assessment of society

in Rostock (top third) than in Jena (middle field), also both Spanish locations

differed with a more positive attitude in Sant Cugat (top third) compared to

Vic (middle field). Estonia is divided into positively answering respondents

from Tartu (top third) contrasting the Narva area (bottom third). The

Croatian field sites are located in the end of the top third (Podsljeme) and in

the end of the middle field (Pescenica). Hungarian respondents revealed an

average assessment of trustworthiness and helpfulness in Sopron and an

under average one in Ozd.

MYPLACE August 2014


10.3 Interaction with friends (Q31)

Question 31 covers the frequency of personal contacts to close friends asked within a time

framework of the last year. Answer options offered ‘Several times a week’, ‘Once a week’,

‘At least once a month’, ‘Several times a year’, ‘Once a year or less’ and ‘I don`t have any

close friends’.

As the following question (Q32) on the duration of media contact to friends refers to a

different time frame (an average weekday and duration instead of frequency), it was

impossible to combine these two questions. Consequently, question Q31 was, handled as a

single variable in the following analyses.

The coding was, reversed to identify the most frequent answer option with the highest

value.

MYPLACE August 2014


Q31: Interaction with friends


The distribution of this variable is left skewed (more respondents met

their friends at least once a week). The graph is, ordered combining the

categories ‘once a week’ and ‘several times a week’.

Narrative

16,821 respondents answered this question. 65% of them met their

friends more than once a week during the last year, another 20% met

their friends once a week and 10% at least once a month. 4% said, they

met their close friends several times a year, 1% reported one or less

meetings per year and 0.5% said that they do not have any close friends.

In New Philadelphia (Greece) the highest share of respondents (97%)

was found meeting their friends at least once a week.

Vyborg (Russia) had the smallest proportion (71%).

In most countries, both locations are near each other in the ranking. One

exemption is Finland as Lieksa and Nurmes belong to the top third, while

Kuopio belongs to the middle field. Bremen has a position in the upper

middle field, whereas the second location from western Germany

(Bremerhaven) is, situated in the upper bottom third. Also, the Danish

field sites can be found in the middle field (Odense East) and in the

bottom section (Odense Center). However, differences are only marginal

between locations from one country. The Mediterranean countries

(Greece, Croatia, Spain) and Portugal are dominating the top third with

Finland and Georgia following.

Graph 10.3: Frequency of meetings with close friends by location

MYPLACE August 2014


10.4 Use of media technologies and social networking sites

(Q32)

For estimating the respondent`s contact to friends via media technologies, they were asked

for the duration of communicating with their friends via social media or other technologies

(e.g. text messages). The following answer options were offered: ‘No time at all’, ‘Less than

1 hour’, ‘1 hour to 2 hours’, ‘More than 2 hours’. This question was handled as a single

variable in the following analyses.

MYPLACE August 2014


Q32: Use of media technologies and social network sites


The distribution of this variable is almost normal, but slightly left skewed

(only few respondents spend ‘no time at all’ with contacting their friends via

media technologies). The graph is, ordered by the category ‘more than 2

hours’.

Narrative

This question was answered by 16,771 respondents. 32% reported that they

contact their friends via media technologies for more than 2 hours during an

average week, 34% for one to two hours and 29% for less than one hour. 5%

said that they spend no time at all with media contacts to friends.

Countries with both locations situated in the same third of social media

usage are UK, Georgia Russia and Croatia (top third), Denmark (middle

section) and Latvia and eastern Germany (bottom third). Estonia and Spain

have locations in the top third (Narva area and Saint Cugat) and in the

middle field (Tartu and Vic). Most pronounced differences between locations

of one country can be found in Portugal, where Lumiar belongs to the top

third and Barreiro to the bottom. In western Germany, Finland, Hungary and

Greece respondents from the two sampled locations either revealed a

medium usage of social media technologies (Bremerhaven, Lieksa and

Nurmes, Ozd, New Philadelphia) or a low level (Bremen, Kuopio, Sopron,

Argyroupouli) of media usage for contacting friends (compared to other

locations).

Graph 10.4: Use of media technologies and social network sites by

location

MYPLACE August 2014


10.5 Diversity of social networks (Q33)

The diversity of respondent`s circle of friends was evaluated with the help of a battery of six

sub questions. Respondents were asked how many of their friends

Q33_1 are members of a different race/ethnic or minority group,

Q33_2 have a different social status/class,

Q33_3 have different political views,

Q33_4 have different religious beliefs,

Q33_5 have a different sexual orientation,

Q33_6 have a different gender.

All questions belonging to Q33 were recoded, in order to give the value 0 (instead of 1) for

the category ‘no differing friends’.

Factor analyses suggested combining all six sub questions into one scale (“diversity of

friends scale”). Good reliability was proofed by Cronbach`s α (0.78).

However, it is important to note that due to many missing responses in Q33_2, Q33_3,

Q33_4 and Q33_5 (see table 10.5), the diversity of friends scale has 22.3% missing values.

Table 10.5 Proportion of Missing Values by question per country

How many of your friends are/have …

Country

Derived

variable

…

members

of a

different

race /

ethnic or

minority

group to

yours?

… a

different

social

status /

class to

yours?

…

different

political

views to

yours?

…

different

religious

beliefs to

yours?

… a

different

sexual

orientation

to yours?

… a different

(sex/gender) than

yours?

Croatia 29 12 12 22 12 13 7

Denmark 10 1 4 5 3 3 0

Estonia 31 8 9 21 16 9 3

Finland 40 7 21 28 14 15 5

Georgia 12 0 2 10 1 2 1

Germany-

western 14 1 6 7 2 3 0

Germany-

eastern 10 0 4 6 2 2 0

Greece 16 1 2 12 3 3 0

Hungary 35 6 7 27 29 14 6

Latvia 11 1 2 6 5 2 2

Portugal 40 3 6 28 20 15 2

MYPLACE August 2014


Russia 28 5 9 18 16 10 4

Slovakia 32 8 9 22 20 12 5

Spain 6 1 2 4 1 1 0

UK 17 1 5 12 2 4 1

Total 22% 4% 7% 15% 10% 7% 2%

MYPLACE August 2014


Q33: Diversity of social networks


The scale’s distribution is almost normal, however slightly left skewed (less

diversity in circle of friends). Values on the diversity of friends scale range

from 0 (homogeneous friend network) to 18 (maximum heterogeneous

friend network).

Graph 10.5: Mean values for diversity of social networks by location

Narrative

The overall mean value for the diversity of friends scale is 8.25 (n=13,155,

sd=4.98). The highest mean score is given in Barreiro (13.09, Portugal) and

the lowest in Vyborg (3.80, Russia).

Portugal, the UK, western Germany and eastern Germany are the countries,

where both locations reveal a mean value of 9.5 or above. Croatia and

Denmark show similar results, however both have one location with a lower

mean value. The Spanish locations are both located in the middle section.

Latvian field sites are also alike and located in the top of the bottom section.

Countries where circles of friends were least diverse in both locations are

Georgia and Hungary.

Most pronounced differences between two locations of one country were,

observed in Greece (7.40 in New Philadelphia; 6.85 in Argyroupouli), Finland

(9.67 in Kuopio; 7.15 in Lieksa and Nurmes), Estonia (8.64 in Tartu; 5.55 in

the Narva area), Slovakia (9.01 in RimavskaSobota; 6.44 in Trnava) and

Russia (7.43 in Kupchino; 3.80 in Vyborg).

MYPLACE August 2014


10.6 Attitude towards roles of women and men in society

(Q34)

Two questions were added in the section “Social networks, gender relationships and

sexuality” to enlighten respondents` attitudes towards women`s role in society. The first

one refers directly to the labour market and the second one to (political) leadership.

Respondents were asked to agree or disagree on a 5 point scale to the following statements:

Q34_1: When jobs are scarce, men should have more right to a job than women.

Q34_2: Women make as good political leaders as men.

Answer options were ‘Strongly agree’, ‘Agree’, ‘Neither agree nor disagree’, ‘Disagree’ and

‘Strongly disagree’.

For the following analyses, the scale for Q34_2 was recoded in order to receive the highest

value for the most positive attitude towards women in both sub questions. Both questions

were added to a scale. However, Cronbach`s α implied a poor reliability (0.54).

Nevertheless, the correlation coefficient calculated was 0.37. Thus, a scale consisting of

Q34_1 and Q34_2 can be used.

MYPLACE August 2014


Q34: Attitudes towards roles of women and men in society (Q34_1 + Q34_2)


The scale, ranging from 0 to 8, is distributed left skewed (more positive

attitudes towards women`s role in society).

Graph 10.6: Mean attitudes of gender roles by location

Narrative

Overall mean is 5.88 (n=16,545, sd=1.78). The highest mean score was

observed for Odense, Denmark (7.19) and the lowest one for Kupchino,

Russia (4.67).

Both locations of Denmark, Spain, Finland and eastern Germany are

characterised by mean values as top third. Portugal and western Germany

also have similar mean values in both locations and belong to the top third

(Lumiar, Bremen) and the middle section (Barreiro, Bremerhaven). A middle

position for both locations can be observed for the UK, Croatia and Latvia.

The lowest mean values for agreement to a positive attitude towards

women`s roles in society was given for both locations of Georgia, Greece,

Slovakia and Russia.

Except for Hungary and Estonia, locations belonging to each country had

close positions in the attitude ranking. Sopron respondents (Hungary) had a

mean value of 6.00, while Ozd respondents had a score of 5.31. In Estonia

Tartu yielded a mean value of 5.83 and the Narva area 4.86.

MYPLACE August 2014


10.7 Attitude towards homosexuality (Q35)

Two questions captured respondent`s attitude towards homosexuality. Respondents were

asked to agree or disagree on a 5 point scale to the following statements:

Q35_1: Homosexuality is unnatural.

Q35_2: Same sex couples should be able to adopt children.

Answer options were ‘Strongly agree’, ‘Agree’, ‘Neither agree nor disagree’, ‘Disagree’ and

‘Strongly disagree’.

For the following analyses, the scale for Q35_2 was recoded in order to receive the highest

value for the most positive attitude towards homosexuality in both sub questions. Both

questions were added to a scale. Cronbach`s α underlined a good reliability (0.77) for the

‘homosexuality scale’. Additionally, a correlation coefficient was calculated (0.62), also

indicating a good reliability. Thus, a scale consisting of Q35_1 and Q35_2 can be used.

MYPLACE August 2014


Q35: Attitude towards homosexuality (Q35_1 + Q35_2)


The homosexuality scale is, almost normally distributed, however slightly

skewed to the left (implying a more positive attitude towards

homosexuality). The scale ranges from 0 to 8.

Graph 10.7: Mean values for attitude towards homosexuality by

location

Narrative

15,834 respondents answered the questions on attitudes towards

homosexuality. The overall mean value is 4.25 (sd=2.52). Highest mean score

is given in Jena (eastern Germany) and lowest, expressing least tolerant level

towards homosexuality, in the Narva area (Estonia).

Most tolerance towards homosexuality for both locations was, expressed in

eastern and western Germany, Denmark, Spain and Finland. One location

from UK (Nuneaton) also belongs to the top third, while the second British

field site (Coventry) is, situated in the middle section. Both locations from

Portugal and Hungary can, be found in the middle segment. Also Russia can

be considered as middle-positioned as Vyborg belongs to the middle section

and Kupchino has the first position of the bottom third. Greece, Latvia and

Georgia are located in the bottom section for mean scores, for tolerance

towards homosexuality with both locations.

Notable differences in the ranked positions can be found for Estonia (Tartu

with 4.00 in the middle section and the Narva area with 2.04 in the bottom),

Croatia (Pescenica with 3.48 in the middle and Podsljeme with 2.58 in the

bottom third) and Slovakia (Rimavska Sobota with 3.36 in the middle and

Trnava with 2.92 in the bottom).

MYPLACE August 2014


10.8 Views on women to have access to abortion (Q36)

Respondents were, asked whether they thought that women should have the right to have

access to abortion or not. Answer options were ‘Yes’ or ‘No’.

As there were no other variables in the section “Social networks, gender relationships and

sexuality” suitable for a combination with this question, Q36 was handled as a single

variable in the following analyses.

MYPLACE August 2014


Q36: Views on women to have access to abortion


The graph is, ordered by the agreement to women to have access to

abortion.

Graph 10.8: Views on women to have access to abortion by location

Narrative

Overall, 15,485 respondents answered this question, with 77% of

respondents stating that women should have access to abortion.

Coventry (UK) had the largest proportion (98%) of agreement to

women`s right to abortion and Kutaisi (Georgia) the lowest (11%).

Apart from Georgia (Telavi: 17%; Kutaisi: 11%) agree to 59% or more

with the statement that women should have the right to abortion.

In the UK, eastern Germany, Finland and Denmark both locations

belong to the top third. Locations in western Germany, Spain,

Portugal and Slovakia mark are, situated in the middle section.

Georgia is, only represented in the bottom third. Greek, Croatian and

Estonian locations are situated in the top third (Argyroupouli,

Pescenica, Tartu) and in the bottom third of agreement values (New

Philadelphia, Podsljeme, Narva area). Russia, Hungary and Latvia have

each one location in the middle section (Vyborg, Sopron, Agenskalns)

and in the bottom section (Kupchino, Ozd, Forstate&Jaunbuve).

MYPLACE August 2014



Dependent variable

The dependent variable used is the social trust scale computed on basis of question Q29 and

Q30 (see above). As this variable is an almost normally distributed scale variable, a linear

mixed model was used.


On the individual level, several socio-demographic independent variables were used.

According to Kelly (2009), factors like a person`s economic situation have a major impact on

social exclusion and thus, on trust in society. Thus, we included variables for income

satisfaction (ordinal from 0=very difficult to live on, 1=difficult to live on, 2=coping to

3=living comfortably) and the respondent`s employment status (nominal from 0=Other,

1=Unemployed, 2=In education to 3=Employed) in the analyses. We also included the social

class variable (ordinal from 0=lowest class to 3=highest class) as a complementation.

Additionally, we used gender (dummy) and age (z-transformed) as control variables.

Assuming that a strong religious belief could have a positive effect when it comes to trusting

other people, we included a dummy in the analyses giving information if the respondent

was religious or not. Another aspect connected with integration in society is ethnicity.

Belonging to a country`s ethnic majority is assumed to convey integration and thus, a

positive attitude towards society. For this reason, we included a dummy variable for ethnic

majority.

Contextual variables used in the multi-level analyses were, extracted from administrative

datasets. One exception, the welfare state type (ordinal, from 0=Post-socialist, 1=Nordic,

2=Conservative, 3=Mediterranean to 4=Liberal), came from the original MYPLACE dataset.

Different welfare state types indicate different levels of social expenditures and thus, are,

expected to have an impact on social inclusion/exclusion, linked with trust in society.

The administrative data used gave information on:

a. Various aspects of the political process, civil liberties and political rights. b. Political stability. c. The quality of public service provision, the quality of the bureaucracy, the

competence of civil servants, the independence of the civil service from political pressures, and the credibility of the government’s commitment to policies.

d. Market regulatory qualities. e. The extent to which, agents have confidence in and abide by the rules of society.

These include perceptions of the incidence of crime, the effectiveness and predictability of the judiciary, and the enforceability of contracts.

f. The level of control of corruptions. g. GDP.

MYPLACE August 2014


All these variables were z-transformed and implying differences on the national level with

respect to economic aspects (d, g) (see for theoretical importance: Kelly, 2009),

opportunities of participation (a) (see for theoretical importance: Cook et al., 2003; Hardin,

2000) and stability and reliability issues (b, c, e, f) (see for theoretical importance:

Nannestad, 2008; Rothstein and Eek, 2009).

Results

The empty model shows an intra-class correlation of 0.17, thus 17 percent of the overall

variance in the social trust scale can be, explained on the aggregate level of localities. This

result indicates that a multi-level approach should be used.

The inclusion of the described individual variables had only marginal impact on the

improvement of the model (loglikelihood decreased by 45.55) and the estimators on the

individual level and the location level (compared to empty model 1.37% more variance was

explained). The overall variance dropped from 17.01% in the empty model to 16.77% in the

model including the individual variables. Thus, the individual variables did not have a major

impact on variance explanation for the social trust scale.

Table 10.9.1. Variance components of regression models explaining trust in society Empty

model

(random

intercept

only)

With

individual

level

explanatory

var’s

With

country

level

explanatory

var’s


Trust in society σ (location level) 2.834 2.666 0.795

-loglikelihood 85615.511 85569.966 85507.774

Expanding the model by the contextual variables reveals some improvements. Although,

loglikelihood does not indicate a major improvement (62.19 difference), the overall variance

drops to 14.90, which can be explained by an improved variance explanation on the location

level only. 12.56% of variance can be explained on the location level in the third model.

Thus, the added contextual variables referring to a country`s economical, political and

stability situation influence people`s trust in society. Table 10.9b depicts the coefficients of

the third model (significant influences printed bold).

MYPLACE August 2014


Table 10.9.2.Coefficients of regression models explaining trust in society (n1=15,572 ;

n2=30)

Among the individual variables included in the model, satisfaction with income, the

respondent`s employment status and social class as well as age have a significant impact on

trust in society. Respondents who are being dissatisfied with their income (p <.000 for the

two lowest categories of satisfaction) are less likely to develop a high level of trust in

society, compared to those who say that they are living comfortably on their income.

Connected with this result, unemployed persons and those coming from the lower middle

class are less likely to trust other people (p < .05 each) compared to, employed ones and

those coming from the upper social class. Surprisingly, the younger the interviewed

respondents were, the less trust in society they revealed (p < .000).

The contextual variables showed that Post-socialist and Mediterranean welfare states

convey trust in society (p < .05 each). Additionally, respondents living in societies with

Model 3

b SE


Gender (woman) -0.109 (0.062)

Ethnic majority (no) -0.100 (0.090)

Income (very difficult to live on) -0.748 *** (0.139)

Income (difficult to live on) -0.426 *** (0.099)

Income (coping) -0.074 (0.079)

Income (living comfortably) Ref. Ref.

Religious (no) -0.125 (0.078)

Employment situation (Other) -0.199 (0.136)

Employment situation (Unemployed) -0.246 * (0.115)

Employment situation (In education) -0.135 (0.083)

Employment situation (Employed) Ref. Ref.

Social class (lowest) -0.140 (0.085)

Social class (lower middle) -0.235 * (0.112)

Social class (upper middle) -0.095 (0.111)

Social class (highest) Ref. Ref.

Age (z-transformed) -0.132 *** (0.036)


Welfare State ( Post-socialist) 6.710 * (3.161)

Welfare State (Nordic) -0.437 (1.281)

Welfare State (Conservative) 2.780 (1.348)

Welfare State (Mediterranean) 4.916 * (2.204)

Welfare State (Liberal) Ref. Ref.

Corruption control index (z-transformed) -3.415 (3.022)

Government effectiveness index (z-transformed) 9.526 (5.979)

Political stability index (z-transformed) -0.963 (1.580)

Rule of law index (z-transformed) 9.139 ** (2.980)

Market regulatory index (z-transformed) -8.801 * (3.758)

Participation index (z-transformed) -3.128 (2.734)

GDP (z-transformed) 0.546 (0.580)

Constant 0.911 (4.462)

***p<0.001, **p<0.01, *p<0.05, ~p<0.1, two-tailed tests.

MYPLACE August 2014


positive beliefs in the juridical system and lower crime rates are more prevalent to develop

a positive attitude towards society (p < .01) and less strict market regulations also convey

trust in society (p < .05) on a significant level.

In general the theoretical assumption that a person`s economic status influences the

development of people`s trust in society can be confirmed. Additionally, age turned out as

an important predictor for trust in society. It is unclear, what the reason for this might be.

On the contextual side, Post-socialist and Mediterranean regimes both turned out as

welfare state systems with potential for mediating a high level of trust in society. This result

is unexpected as social expenditures are lower in these states than in Nordic or conservative

welfare states. However, the reference category was liberal states, where social

expenditures are also low. The most influencing variables coming from the administrative

dataset were those indices depicting beliefs in the juridicial system and market regulations.

This result suggests that countries without governmental interventions in the economy and

a well-developed juridicial system provide the best conditions for a society with a high level

of trust.

10.10 Summary

The scale expressing the level of trust in society indicated highest levels of trust in Finnish,

western German and Portuguese locations, while Greece, Russia, Slovakia and Georgia had

the lowest mean values for level of trust in society. This scale was also used as a dependent

variable in the multi-level model, which identified respondents from Mediterranean

countries as more likely to develop a high level of trust in society. However, Greece was

located in the bottom third of all participating countries when it comes to mean trust

scores. As the survey was conducted in a period when Greece was hit particularly hard by

the Euro crisis and the stability of the government and market were undermined the low

level of trust measured is comprehensible.

When it comes to the frequency of meeting friends in person, the Mediterranean countries

as well as Portugal had the highest rates. However, in all locations the majority reported

meeting their friends at least once a week.

Media technologies were used with higher duration in the UK, Georgia, Russia and Croatia,

while Latvian and eastern German field sites had more respondents using technologies for

shorter periods. However, the survey asked for the duration of technology usage for social

contacts, not for frequency, which might produce different results.

The question battery Q34 was added to a scale indicating the diversity of respondents’

network of friends. Portugal, the UK and Germany (eastern and western) had the highest

MYPLACE August 2014


mean values. The UK and Germany have a high share of migrants, this might be one reason

for more heterogeneous social networks. Georgian and Hungarian respondents had the

most homogeneous circles of friends.

The most positive attitudes towards women`s role in society and homosexuals were

expressed in the Nordic countries (Denmark, Finland) as well as Spain and Germany.

In Georgia, Greece, Slovakia and Russia, gender based inequality tends to be supported by

the interviewed youths. Rejecting equality for homosexuals was, most often expressed in

Greece, Latvia and Georgia.

Women`s access to abortion is most supported in eastern Germany, Spain and Estonia.

Young people in locations in Russia, Greece and Georgia were the most likely to disapprove

of women’s access to abortion.

The multi-level model testing influences on trust in society revealed that 17% of variance

could be explained at the level of locations. Thus, the context variables were of special

interest. In sum, a country`s welfare state system, the rigidity of the regulation of its market

and juridicial system are decisive aspects when it comes to shaping people`s trust in society.

Additionally, individual economic aspects as well as age play an important role.

References

Cook, K. S. (2003).Trust in Society. New York: Sage.

Collier, K. L., Bos, H. M. W., and Sanford, T. G. M. (2010). Intergroup Contact, Attitudes

Toward Homosexuality, and the Role of Acceptance of Gender Non-conformity in Young

Adolescents. Journal of Adolescence 35(4) pp. 899–907.

Granovetter, M. S. (1973).The Strength of Weak Ties. American Journal of Sociology, 78(6):

1360-1380.

Hampton, K., Sessions Goulet, L., Rainie, L., and Purcell, K. (2011). Social Networking Sites

and Our Lives: How People’s Trust, Personal Relationships, and Civic and Political

Involvement are Connected to Their Use of Social Networking Sites and Other Technologies.

Pew Research Center’s Internet and American Life Project: URL:

http://pewinternet.org/Reports/2011/Technology-and-social-networks.aspx.

Hardin, R. (2000).Trust and Society. In:Galeotti, G., Salmon, P., and Wintrobe, R. (Eds.).

Competition and Structure: The Political Economy of Collective Decisions. Essays in Honor of

Albert Breton(17-46). Cambridge: University of Cambridge Press.

MYPLACE August 2014


Heinze, J. E., andHorn, S. S. (2009).Intergroup Contact and Beliefs about Homosexuality in

Adolescence. Journal of Youth and Adolescence 38: 937–951.

Herek, G. M. (2009). Sexual Prejudice. In:Nelson, T.D. (Ed.).Handbook of Prejudice,

Stereotyping, and Discrimination (441-468).New York: Taylor and Francis.

Kelly, D. C. (2009). In Preparation for Adulthood: Exploring Civic Participation and Social

Trust Among Young Minorities. Youth and Society40(4):526-540.

McPherson, M., Smith-Lovin, L., and Cook, J. M. (2001).Birds of a Feather: Homophily in

Social Networks. Annual Review of Sociology27:415-444.

Misztal, B.A. (1996). Trust in Modern Societies. Oxford: Blackwell.

Nannestad, P. (2008). What have we learned about generalized trust, if anything? Annual

Review of Political Science11: 413–436.

Newman, L. L., and Dale, A. (2005). Network Structure, Diversity, and Proactive Resilience

Building: A Response to Tompkins and Adger. Ecology and Society 10(1): r2. [online] URL:

http://www.ecologyandsociety.org/vol10/iss1/resp2.

Newton, K. (2001).Trust, Social Capital, Civil Society, and Democracy. International Political

Science Review, 22(2): pp. 201-214.

O'Neill, B., and Gidengil, E. (2006).Gender and Social Capital. New York: Taylor and Francis.

Rivers, I., and Duncan, N. (2013).Bullying: Experiences and Discourses of Sexuality and

Gender.New York: Routledge.

Rothstein, B., and Eek, D. (2009).Political Corruption and Social Trust. Rationality and Society

21(1):81–112.

Tao, R., Young, D. L., Ming, L., and Xi, L. (2014). How Does Political Trust Affect Social Trust?

An Analysis of Survey Data from Rural China using an Instrumental Variables Approach.

International Political Science Review35(2): 237–253.

Wimmer, A., and Lewis, K. (2010).Beyond and Below Racial Homophily: ERG Models of a

Friendship Network Documented on Facebook. American Journal of Sociology116(2):583–

642.

Woolcock, M. (2001). The Place of Social Capital in Understanding Social and Economic

Outcomes. Canadian Journal of Policy Research2(1):11–17.

MYPLACE August 2014


Chapter 11: Religion

In this chapter we present the results from the analysis of respondents’ attitudes towards

religion. This was measured using a range of questions covering the degree of religiosity, the

frequency of attending religious services, as well as attitudes towards specific issues

including religion’s role in politics.

Literature

Human religiosity and its different components has been the subject of numerous studies

especially in the western (Küçükcan, 2010; Verbit, 1970). Discussions about the extent to

which religion has declined and society secularized (Kepel, 1991) are central to debates

about religion and have been since at least the 1960s. Young people and religion has been

a separate interest among researchers, e.g. The National Study of Religion and Youth

(www.youthandreligion.org) in the USA. Some of these studies tend to reach the conclusion

that young people are less and less religious and that many of them are self-identified as

atheists, agnostics or not affiliated with any specific religion (Bengtson, 2013; Collins-Mayo,

2012). Furthermore, many studies have explored the association between religiosity and

youth wellbeing and whether religious young people are more or less optimistic than non-

religious ones; but not enough studies were traced concerning the hypotheses of our

research.

Religion and gender is one of the main topics in the fields of the sociology of religion

(Nason-Clark and Neitz, 2001; Hawthorne, 2011), theology (Clark, 1997; Bendroth, 1993)

and feminist studies (Juschka, 2001; Fulkerson, 1994). A significant number of studies have

concluded that women, compared to men, are more religious and are more likely to be

engaged in religious practices (Loewenthal et al., 2002; Woodhead, 2007). It is argued that

women are more likely to participate in social networks that act and are supported by

religious organisations (Krause et al., 2002). Loewenthal et al. (2002) argue that women are

prone to be engaged in religious practices due to socialisation processes that reproduce

specific stereotypes of women, as more caring and sensitive, destined to be in help of

others. Being mothers they spend more time in raising their children than men and part of

this raising is also the religious socialisation.

The same applies to socialisation and nurturing of boys often based on stereotypes

regarding their behaviour in society. What is important to notice is that gender differences

regarding religiosity, are differentiated depending on the religion. For example, Muslim

http://www.youthandreligion.org/

MYPLACE August 2014


women are less active in their religious duties compared to men and this must be analysed

taking into consideration the definition of gender roles in each religious context.

Social class, measured by socioeconomic status, is associated with an individual’s religious

affiliations and practices. But this affiliation has more to do with how religion is practiced

rather than with the degree of religiosity. However, it is often assumed that people who

tend to be critical about religion, or atheists, agnostics and non-religious are members of

the middle or upper classes. Furthermore, it is argued that members of the middle class

tend to belong to more mainstream churches and that members of lower classes tend to be

affiliated with more fundamentalist religions and sect-like groups. Social class is, however,

not significantly correlated to religiosity. The variable of social class could not function as an

index of how strongly religious a person is. Lower classes are more likely than higher ones to

pray in private, believe in the doctrines of their faith, and have intense religious experiences

(Demerath, 1965; Davidson, 1977); the middle and upper classes are more likely to attend

worship services and take part in church organisations and activities. Stark (1972) argues

that the poor show greater religiousness in those aspects of faith that serve as a relief for

suffering; the middle and upper classes participate in religious activities that help confirm

the legitimacy of their claim to high status.

Migrants are believed to be more religious than the native populations and that their

religious behaviour is significantly different from that of the native population, especially

when the religion, they practice does not belong to one of the major religions of the host

country. Williams (1988) states that migrants are more religious in the host country than

they were in their native country. The religiosity of migrants might serve either as factor

that makes their need of assimilation in the new country easier to be fulfilled (Herberg,

1960; Gordon, 1968; Smith, 1978) or as an obstacle that alienates them from the native

community (Greeley, 1972; Cadge and Ecklund, 2006). As it is argued, especially in the USA,

religious institutions of migrants provide social and physical space and social networks that

help the migrants reproduce and maintain their values traditions and customs in the midst

of an often alienating and strange society. In that sense, religion becomes a way of

reproducing many aspects of an immigrant’s native culture through collective memory and

symbolic rituals (Ebaugh 2003:230; Ebaugh and Chafetz, 2000).

11.1 Religiosity (Q37)

Religiosity is measured using an eleven-point scale where high values indicate high level of

religiosity. It is made clear to the respondents that religiosity is examined regardless of the

religion.

MYPLACE August 2014


Q37: Religiosity


The distribution of the responses is normally distributed and provides

a good eleven-point scale, where high values indicate strong

religiosity.

Graph 11.1: Mean religiosity by location

Narrative

The overall mean for religiosity is 4.48 (n=16,722, sd=3.169). This

varies by location, with the highest mean score in Telavi (Georgia)

with 8.19, and the lowest mean score in Odense Center (Denmark)

with 2.43.

There are clear clusters of locations; locations in Georgia, Greece,

Croatia and Russia all in the top third, with higher levels of religiosity.

Eastern Germany, Spain and Denmark are clustered in the bottom

third with lower levels of religiosity.

Most countries have similar mean score between the two locations;

notable exceptions are; UK (Coventry and Nuneaton), 1.94 difference

between two mean scores and Estonia (Narva area and Tartu), 1.67

difference.

8,19

8,10

5,73

5,69

5,57

5,49

5,48

5,48

5,27

5,22

5,14

4,78

4,65

4,28

4,10

4,07

3,82

3,80

3,79

3,61

3,48

3,41

3,35

3,29

3,25

3,19

2,92

2,75

2,61

2,43

,00 1,00 2,00 3,00 4,00 5,00 6,00 7,00 8,00 9,00

Telavi (GEO)

Kutaisi (GEO)


Podsljeme (CRO)

Vyborg (RUS)

Kupchino (RUS)


Argyroupouli (GRE)

Pescenica (CRO)

Coventry (UK)

Trnava (SLO)


Sopron (HUN)

Narva area (EST)

Bremen (GER-W)

Agenskalns (LAV)

Barreiro (POR)

Bremerhaven (GER-W)

Lumiar (POR)

Kuopio (FIN)

Ozd (HUN)

Odense East (DEN)


Nuneaton (UK)

Vic (SPA)

Jena (GER-E)

Rostock (GER-E)

Sant Cugat (SPA)

Tartu (EST)

Odense Center (DEN)

MYPLACE August 2014


11.2 Attending religious services (Q38)

Frequency of attending religious services is investigated using the following scale:

Many times/ week

Once/ week

At least once/ month

More than twice/ year

Once or twice/ year

Less than once/ year

Never

An important clarification for the respondents is that the frequency examined does not

include weddings, baptisms and funerals.

MYPLACE August 2014


Q38: Frequency of attending religious services


The ‘many times/ week’ and ‘once/ week’ responses are unified in the

graph and labelled as ‘at least once/week’. The graph is ordered by

combining the ‘at least once/week’ and the ‘at least once/month’

responses.

Graph 11.2: Frequency of attending religious services by location

Narrative

The highest frequency of attending religious services appears in

Kutaisi (Georgia), where 45% of the respondents attend at least once/

week, as well as Telavi (Georgia), where the corresponding figure is

36%. The lowest frequency appears in Odense Centre (Denmark),

where 2% of the respondents attend services at least once/ week and

38% never attend.

There are clear clusters of locations; with locations in Georgia, Croatia

and Greece all in the top third, with higher frequency of attending

religious services. Locations in Finland and Estonia are clustered in the

bottom third.

45

36

28

26

20

20

9

10

7

12

5

11

12

8

7

8

7

10

8

9

5

6

4

4

3

4

6

2

2

2

41

38

15

14

17

12

18

12

14

8

15

8

6

8

8

6

8

3

5

4

8

5

7

5

5

4

1

3

2

1

13

23

26

40

29

39

49

37

51

25

34

18

19

35

37

39

47

41

22

26

35

17

38

34

30

41

44

39

31

43

1

1

9

8

10

10

12

26

16

14

19

9

10

19

23

10

20

13

8

8

11

22

13

12

23

24

23

27

23

17

1

1

21

11

24

18

12

15

13

41

27

54

53

30

25

36

19

33

57

53

42

50

38

45

38

27

26

29

41

38

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Kutaisi (GEO)

Telavi (GEO)

Coventry (UK)

Podsljeme (CRO)

Trnava (SLO)

Pescenica (CRO)


Vyborg (RUS)

Argyroupouli (GRE)

Lumiar (POR)

Sopron (HUN)

Barreiro (POR)

Nuneaton (UK)


Kupchino (RUS)

Bremen (GER-W)


Odense East (DEN)

Sant Cugat (SPA)

Vic (SPA)

Bremerhaven (GER-W)

Ozd (HUN)

Jena (GER-E)

Rostock (GER-E)

Narva area (EST)

Agenskalns (LAV)

Kuopio (FIN)


Tartu (EST)

Odense Center (DEN)

at least once a week at least once a month at least once a year

less than once a year never

MYPLACE August 2014


11.3 Attitude towards religion and the role of religion in

politics (Q39)

Respondents’ attitudes towards religion and its role in politics were measured by a four-

item question. Two of the items examine attitudes towards religious truths (items 1 and 4)

and two of the items investigate the attitudes toward religion’s role in politics (items 2 and

3).

Q39_1 There is only one true religion

Q39_2 Good politicians believe in God

Q39_3 Religious leaders should have no influence on government decisions

Q39_4 None of the major religions have any truths to offer

The factor analysis of these four items did not produce distinct factors, therefore we did not

produce indices and analysed each of the items separately.

MYPLACE August 2014


Q39_1: There is only one true religion


The graph is ordered by combining the ‘strongly agree’ and the ‘agree’

responses.

Graph 11.3a: There is only one true religion

Narrative


respondents who strongly agree, 16% agree, 18% neither agree nor

disagree, 27% disagree and 29% strongly disagree, that there is only

one true religion.

Kutaisi (Georgia) had the largest proportion of respondents (89%)

strongly agreeing or agreeing. Rostock (eastern Germany) had the

smallest proportion (2%).

Rimavska Sobota (Slovakia) had the largest proportion of respondents

(35%) who neither agree nor disagree.

Georgia and Greece are the countries with overall higher levels of

agreement, as opposed to eastern Germany that uniformly has the

lowest levels of agreement.

44

40

21

18

12

17

14

14

8

7

10

7

12

9

4

5

9

10

6

4

9

7

5

4

5

3

3

2

2

1

45

41

29

31

28

22

23

20

21

21

15

18

13

16

19

17

11

8

11

12

7

6

6

5

3

5

4

2

1

1

4

8

17

22

25

16

31

24

27

35

21

32

29

24

23

19

11

10

21

26

18

24

8

7

11

9

8

10

8

9

7

11

21

17

18

26

23

24

28

23

31

31

31

41

38

41

33

27

39

30

31

16

32

32

13

33

45

25

17

18

0

1

12

12

17

20

9

17

15

15

22

12

15

11

16

18

35

45

23

28

36

47

48

52

68

50

40

62

72

71

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Kutaisi (GEO)

Telavi (GEO)


Argyroupouli (GRE)

Trnava (SLO)

Coventry (UK)

Vyborg (RUS)

Podsljeme (CRO)

Ozd (HUN)


Pescenica (CRO)


Kupchino (RUS)

Narva area (EST)

Barreiro (POR)

Agenskalns (LAV)

Vic (SPA)

Odense East (DEN)

Nuneaton (UK)

Lumiar (POR)

Sopron (HUN)


Sant Cugat (SPA)

Bremen (GER-W)

Kuopio (FIN)

Bremerhaven (GER-W)

Tartu (EST)

Odense Center (DEN)

Jena (GER-E)

Rostock (GER-E)

always often sometimes rarely never

MYPLACE August 2014


Q39_2: Good politicians believe in God



responses.

Graph 11.3b: Good politicians believe in God

Narrative



disagree, 28% disagree and 27% strongly disagree, that good

politicians believe in God.

Nea Philadelphia (Greece) had the largest proportion of respondents

(38%) strongly agreeing or agreeing. Vyborg (Russia) had the smallest

proportion (1%).

Telavi (Georgia) had the largest proportion of respondents (45%) who

neither agree nor disagree.

There are less clear clusters of locations, with increased uniformity of

results appearing in Croatia, Spain, western Germany and Hungary.

11

7

11

9

7

3

8

8

2

4

1

3

5

3

4

2

2

3

3

4

1

4

1

2

1

0

1

1

0

0

27

26

20

20

22

24

17

16

21

19

21

18

16

17

16

16

15

13

12

7

9

4

6

5

4

4

3

3

3

1

19

32

41

34

45

27

31

30

30

27

25

39

37

35

35

31

27

31

35

33

39

33

19

18

18

22

14

14

18

15

22

33

26

21

16

35

23

35

34

24

28

26

28

26

18

39

26

27

33

30

34

16

38

29

44

14

37

21

23

20

22

2

3

17

11

11

21

11

12

26

25

14

13

19

27

12

30

26

18

26

18

43

36

47

33

61

45

63

56

64

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%


Telavi (GEO)

Kutaisi (GEO)

Argyroupouli (GRE)

Vyborg (RUS)

Barreiro (POR)

Podsljeme (CRO)

Narva area (EST)


Trnava (SLO)

Bremerhaven (GER-W)

Coventry (UK)

Kupchino (RUS)

Ozd (HUN)

Sopron (HUN)

Agenskalns (LAV)

Bremen (GER-W)

Lumiar (POR)


Pescenica (CRO)

Nuneaton (UK)


Vic (SPA)

Odense East (DEN)

Tartu (EST)

Kuopio (FIN)

Sant Cugat (SPA)

Jena (GER-E)

Rostock (GER-E)

Odense Center (DEN)


MYPLACE August 2014


Q39_3: Religious leaders should have no influence on government decisions


The graph is ordered by combining the ‘strongly agree’ and the


Graph 11.3c: Religious leaders should have no influence on

government decisions

Narrative



disagree, 14% disagree and 4% strongly disagree, that religious

leaders should have no influence on government decisions.

Sant Cugat (Spain) had the largest proportion of respondents (83%)

strongly agreeing or agreeing. Nuneaton (UK) had the smallest

proportion (34%).

Vyborg (Russia) had the largest proportion of respondents (45%)

who neither agree nor disagree.

There are clear clusters of locations, with locations in Spain, Greece

and western Germany all in the top third, agreeing that religious

leaders should have no influence on government decisions. When

ordered by ‘disagree’ and ‘strongly disagree’, locations in the UK are

in the top third, disagreeing with the statement.

49

40

25

33

22

36

36

30

26

33

20

19

31

25

29

18

31

28

13

22

19

22

11

27

11

15

11

15

14

12

35

38

46

35

45

28

28

34

37

30

42

43

29

34

30

40

27

30

44

34

35

29

39

22

35

31

33

28

27

21

7

10

17

21

22

18

21

23

25

22

19

31

26

27

20

22

22

29

28

24

35

27

28

31

17

37

22

24

45

35

8

10

9

9

9

13

12

9

9

10

15

6

9

11

16

16

14

10

13

17

7

14

15

14

33

10

30

27

11

27

2

1

3

2

2

6

4

5

3

6

4

1

6

4

5

4

5

3

3

3

3

9

8

6

3

7

3

7

4

4

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Sant Cugat (SPA)

Vic (SPA)

Bremen (GER-W)

Pescenica (CRO)

Tartu (EST)

Odense Center (DEN)

Jena (GER-E)

Argyroupouli (GRE)


Kuopio (FIN)

Bremerhaven (GER-W)


Sopron (HUN)

Kupchino (RUS)

Odense East (DEN)

Barreiro (POR)

Rostock (GER-E)

Podsljeme (CRO)

Agenskalns (LAV)

Lumiar (POR)

Ozd (HUN)

Trnava (SLO)

Narva area (EST)


Telavi (GEO)


Kutaisi (GEO)

Coventry (UK)

Vyborg (RUS)

Nuneaton (UK)


MYPLACE August 2014


Q39_4: None of the major religions have any truths to offer


The graph is ordered by combining the ‘strongly agree’ and the


Graph 11.3d: None of the major religions have any truths to offer

Narrative



disagree, 30% disagree and 11% strongly disagree, that ‘None of the

major religions have any truths to offer’.


strongly agreeing or agreeing. Telavi (Georgia) had the smallest

proportion (4%).

Odense Centre (Denmark) and Rimavska Sobota (Slovakia) had the

largest proportion of respondents (45%) who neither agree nor

disagree.

There are clear clusters of locations, with locations in Spain, Finland

and Latvia in the top third, agreeing that none of the major religions

have any truths to offer. When ordered by ‘disagree’ and ‘strongly

disagree’, locations in Georgia and the UK are clustered in the top

third, disagreeing with the statement that none of the major

religions have any truths to offer.

23

19

31

15

9

18

9

11

15

16

6

12

10

9

6

9

9

7

5

9

8

11

5

5

5

4

3

2

2

1

30

31

15

28

30

20

29

25

19

17

22

16

18

15

16

13

13

14

16

12

12

9

14

13

13

11

11

9

8

3

20

24

38

33

42

27

38

34

34

38

35

45

36

42

40

36

45

29

37

37

26

41

29

34

26

41

26

36

10

8

22

22

8

18

17

23

20

24

19

20

25

20

24

23

26

29

23

34

35

31

36

25

38

34

39

35

40

26

65

69

6

4

7

5

1

11

4

5

13

9

12

7

12

11

11

13

9

15

7

10

17

15

14

14

16

10

20

27

15

19

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Sant Cugat (SPA)

Vic (SPA)


Lumiar (POR)


Kuopio (FIN)

Agenskalns (LAV)

Barreiro (POR)

Trnava (SLO)

Kupchino (RUS)

Narva area (EST)

Odense Center (DEN)

Vyborg (RUS)

Odense East (DEN)

Ozd (HUN)

Argyroupouli (GRE)


Jena (GER-E)

Tartu (EST)

Pescenica (CRO)


Podsljeme (CRO)

Bremerhaven (GER-W)

Rostock (GER-E)

Bremen (GER-W)

Nuneaton (UK)

Coventry (UK)

Sopron (HUN)

Kutaisi (GEO)

Telavi (GEO)


MYPLACE August 2014



The determinants of “religiousness”

Dependent variable

Our dependent variable is “religiousness”, measured in MYPLACE survey by an eleven-scale

question (Q37): “Regardless of whether or not you belong to a particular religion, how

religious would you say you are? 0 means not religious at all and 10 very religious”.


Having reviewed literature on religion and youth, a number of standard socio-economic

variables were included in the model that explain religiousness:

Gender, on the basis of substantial research conducted until today worldwide, investigating possible gender differences regarding religiosity, as well as participation on religious activities.

Parental social class, based on the assumption that lower social status increases the degree of religiousness. Parental is preferred instead of personal status since many young people have not entered the labor market yet and their living standards heavily rely on their parent’s socio/economic status.

Respondent’s economic activity, based on the same rationale explained above: unemployed individuals might express higher need for religiousness, whereas employed – therefore more independent - young people are potentially less religious.

Religion, deriving from the hypothesis that certain religions such as Muslim and Jewish, are associated with more intense religiosity compared to others.

Identity, since being part of the minority in a country might increase the need for religiousness, in the sense of “belonging” to a group as well as in order to feel empowered.

Additionally, a context variable was incorporated in the model, the welfare state types, in

order to investigate whether living in a post-communist country has played a role in shaping

religiousness.

Results

The initial analysis of variance components in the ‘empty’ model in Table 11.8a reveals that

80.3% of the variance in young people’s ‘religiousness’ can be explained by differences

between individuals. Furthermore, an intra-class correlation of 0.197 suggests that 19.7

percent of the overall variance in ‘religiousness’ lies at the level of localities. The substantial

variation at local level justifies the reason for using multilevel analysis in our data.

The inclusion of the described individual variables (in Model 1 in Table 11.8a) resulted in a

27.0% decrease of residual, individual level variance. Chi-square test of deviances

(loglikelihood = 4926.816; df = 13) confirms that this is a significant difference. Adding

individual level variables, led to a 44.6% reduction in the intercept variance, indicating that

MYPLACE August 2014


the amount of variance of ‘satisfaction’ with democracy between localities can be attributed

to individual level predictors.

In the next model (Model 2 in Table 11.4.1), we included contextual variables which led to a

75.4% reduction of intercept variance, compared to the model containing only individual

level predictors. Chi-square test of deviances (loglikelihood = 5287.13; df = 4) confirms that

this is a significant difference. The ICC value of .054 in the final model indicates that 5.4% of

the unexplained variance remains at the level of localities. In other words, there is a scope

for improving the model by including other relevant variables in the analysis.

Table 11.4.1. Variance components of regression models explaining Young peoples’

religiousness Empty model

(random

intercept only) -

Model 0

With individual

level

explanatory

var’s – Model 1

With country

level

explanatory

var’s – Model 2

Religiousness


σ (location level) 1.9647 1.0884 0.2673

-2 log likelihood 77594.363 72667.547 67380.412

df 3 16 20

Table 11.4.2 illustrates the coefficients of Model 2.

Table 11.4.2: Determinants of Religiousness; n2=20, n1=15,740

Model 2

Individual-level predictors b

SE

Unemployed -0.299 *** 0.059

Employed Ref. Ref.

Gender (male) -0.174 *** 0.035

Female Ref. Ref.

In Minority 0.121 * 0.056

In Majority Ref. Ref.

Catholic 0.476 *** 0.056

Protestant 0.537 *** 0.063

Christian Orthodox 0.514 *** 0.083

Muslim 0.628 *** 0.114

Jewish 0.33 0.454

Other religion 0.082 0.092

None Ref. Ref.

Social Class 0 - Low -0.217 *** 0.048

Social Class 1 -0.11 0.064

Social Class 2 -0.119 0.064

Social Class 3 - High Ref. Ref.

Age -0.041 *** 0.006


Nordic 2.071 *** 0.303

Conservative 1.647 *** 0.3

Mediterranean -0.884 0.256

MYPLACE August 2014


Liberal 0.664 0.4

Post-socialist Ref. Ref.

Constant 4.625 *** 0.152

***p<0.001, **p<0.01, *p<0.05; two-tailed tests.

There are a number of individual level predictors, which are significant in the multi-level

model of religiousness. As far as gender is concerned, females appear more religious than

men, confirming the extensive relevant research findings. Young people who are

unemployed are less religious than those who are employed. Social class is also significant,

with young people from a higher parental social class are more religious than those from

lower social class groups.

In terms of religion, Muslims express higher levels of religiousness, as expected. Protestants

follow, along with Christian Orthodox and Catholics when compared with nonbelievers.

Jewish and Other religions were not significant. The age of young people was significant,

with older respondents stating they were less religious than younger respondents were. This

could be due to the greater influence of home life/parents over younger people regarding

religion, and as young people become more independent, they become less religious.

Moreover, being a minority member increases religiousness (p<.05), compared to belonging

in the majority within the country. This finding is in line with our initial hypothesis. We

expected that people who don’t belong to the mass culture express their faith more

intensely for a number of reasons. To begin with, being “different” generates the need to

strengthen one’s identity; religion can become one of the fundamental components of

personal identity. Moreover, being religious can mean attending religious services,

preserving traditions and executing rituals, all the above potentially contributing to social

integration. Furthermore, being a minority could lead to marginalization, thus increasing the

need to feel empowered – a possible benefit of intense religiousness.

Welfare State type contextual predictors, illustrate that young people from Nordic and

Conservative welfare state types have significantly higher levels of religiousness than post-

socialist types. Mediterranean countries are less religious, but are not statistically

significant.

11.5 Summary

Measuring religiosity is a difficult task due to the variety of components that it includes. In

the present survey we used an eleven-point scale where respondents categorised

themselves based on whether they believed they were religious or not. From the

comparative analysis of results, different clusters of countries appeared according to their

MYPLACE August 2014


level of religiosity. Georgia, Greece, Croatia and Russia were part of the cluster which

exhibited high scores on religiosity, whereas eastern Germany, Spain and Denmark

constituted the cluster with low levels of religiosity.

Regarding the frequency of attending religious services, Georgia, Croatia and Greece were

again those countries characterised by high levels of frequency of attending religious

services, whereas Finland, Estonia and Denmark belonged to the cluster exhibiting low

levels of frequency.

Georgia and Greece were the countries in which, overall, higher levels of agreement with

the statement that there is only one true religion were encountered; eastern Germany

demonstrated the lowest levels of agreement. Along the same lines, New Philadelphia in

Greece scored highly in ‘strongly agreeing/agreeing’ with the the statement that good

politicians believe in God. The results of the analysis of responses to that statement showed

that there was significant differences both within and between countries. Eighty three per-

cent of the respondents from Sant Cugat, Spain, ‘strongly agreed or agreed’ that religious

leaders should have no influence on government decisions, whereas 34% of the

respondents in Nuneaton, UK ‘strongly agreed or agreed’ to the aforementioned statement.

Again, three clear clusters of locations were identified, with locations in Spain, Greece and

western Germany agreeing that religious leaders should have no influence on government

decisions. Locations in the UK were in the top third disagreeing with the statement.

Fifty three per-cent of the respondents in Sant Cugat in Spain, ‘strongly agreed or disagreed’

with the statement that none of the major religions have any truths to offer while only 4% of

the respondents in Telavi in Georgia ‘strongly agreed or agreed’ with this statement. Spain,

Finland and Latvia constituted the cluster of countries that agreed more with that

statement, whereas Georgia and UK presented high scores of disagreement.

Regarding the most significant determinants of religiosity, the analysis verified our original

hypotheses. More specifically, gender turned out to be a variable that affects levels of

religiosity, as women appeared more religious than men. Numerous surveys have

repeatedly argued that women are characterised by high levels of religiosity and this

explains the fact that in every piece of quantitative research the gender variable is included

(Stark, 2002:496).

As expected and stated in our original hypotheses, Muslim respondents appear more

religious. Orthodox Christians follow, along with Catholic and Jewish respondents.

Protestant respondents expressed religiousness to a somewhat lesser extent than others

declaring belonging to a faith group, although their religiosity was still significant.

MYPLACE August 2014


Finally, being a part of a minority group increases one’s level of religiosity. This mainly

applies to immigrants who are equipped with different social and cultural capital compared

to the native population and thus try to preserve it and to communicate it to the next

generations in order to maintain their collective memory and identity. Religion in this

process of assimilation plays a crucial role and is vital in the process of crafting an identity.

Religion is not a trait or a characteristic that is strongly associated with a country or a certain

place, and thus helps the people in a country to overcome isolation and marginalisation.

Furthermore being active in religious activities and organisations enables the process of

socialisation and provides people with opportunities for civic engagement and social action.

Finally, when religion is a prominent component of a person’s identity and much of

everyday life is organised around these needs, then it is possible that other differences that

mark an individual’s marginalised place in society, are less pronounced.

References

Bendroth, M.L. (1993). Fundamentalism and Gender: 1875 to the present, New Haven: Yale

University Press.

Bengtson, V.L., Putney, N., Harris, S. (2013). Families and Faith: How Religion is Passed Dawn

Across Generations, New York: Oxford University Press.

Cadge, W. and Ecklund, E.H. (2006). ‘Religious Service Attendance among Immigrants:

Evidence from the New Immigrant Survey-Pilot’, American Behavioral Scientist, No 49 (11),

pp.1574-1595.

Clark, E.A. (ed.) (1997). Women and Religion: The Original Sourcebook of Women in Christian

Thought, San Francisco: Harper.

Collins-Mayo, S. (2012). “Youth and Religion: An International Perspective”, Theo-Web.

Zeitschrift fur Religionspadagogik 11, H.1, pp:80-94.

Davidson, J.D. (1977). “Socio-Economic Status and Ten Dimensions of Religious

Commitment,” Sociology and Social Research, No 61, pp: 462-485.

Demereth, N.J. III (1965). Social Class in American Protestantism, Chicago: Rand McNally.

Ebaugh, H.R. and Chafetz, J.S. (2000). Religion and the New Immigrants: Continuities and

Adaptations in Immigrant Congregations, Walnut Creek, CA: Alta Mira Press.

Ebaugh, H.R. (2003). “Religion and the New Immigrants”, in M.Dillon (ed.), Handbook of the

Sociology of Religion, New York: Cambridge University Press, pp:225-239.

Fulkerson, M. (1994). Changing the Subject: Women’s discourses and Feminist Theology,

Minneapolis: Fortress Press.

Gordon, M.M. (1964). Assimilation in American Life, New York: Oxford University Press.

MYPLACE August 2014


Greeley, A.M. (1972). The Denominational Society: A Sociological Approach to Religion in

America, Glenview, IL: Scott, Foresman.

Hawthorne, S. (2011). “Religion and Gender”, in P.Clarke (ed.), The Oxford Handbook of the

Sociology of Religion, Oxford: Oxford University Press, pp:134-151.

Juschka, D.M. (ed.) (2001). Feminism in the Study of Religion: A Reader, London, New York:

Continuum.

Kepel, G. (1991). La Revanche de Dieu, Paris: Editions du Seuil

Krause, N., Ellison, C.G. and Markum, J.P. (2002). “The effects of church-based emotional

support on health. Do they vary by gender?”, Sociology of Religion, No 63(1), pp: 21-47.

Küçükcan, T. (2010). Multidimensional Approach to Religion: a way of looking at religious

phenomena. Journal for the Study of Religions and Ideologies, No 4(10), pp: 60-70.

Lowenthal, K.M., MacLeod, A.K. and Cinnirella, M. (2002). “Are women more religious than

men? Gender differences in religion activities among different religion groups in UK”.

Personality and Individual differences, No 32, pp: 133-139.

Nason-Clark, N. and Neitz, M. (eds.) (2001). Feminist Narratives and the Sociology of

Religion, Oxford: AltaMira Press.

Smith, T. (1978). ‘Religion and Ethnicity in America’, American Historical Review, No 83, pp:

1155-1185.

Stark, R. (1972). “The Economics of Piety,” in G. Thielbar and S. Feldman (eds) Issues in

Social Inequality. Boston: Little Brown, pp: 483-503.

Stark. R. (2002). Physiology and faith: Addressing the ‘universal’ gender difference in

religious commitment”. Journal for the scientific study of religion, No 41(3), pp: 495-507.

Verbit, M. F. (1970). “The components and dimensions of religious behavior: Toward a

reconceptualization of religiosity”. American mosaic, pp: 24 - 39.

Woodhead, L. (2007). “Gender Differences in Religious Practice and Significance”, in

J.Beckford and N.J.Demerath III (eds.) The Sage Handbook of the Sociology of Religion,

London: Sage, pp: 566-586.

MYPLACE August 2014


Chapter 12: Attitudes towards minority groups

Since the recent global economic crisis, right wing extremism and support for populist

radical right parties across Europe, has been on a rise, driven by rising anti-immigrant and

anti-establishment sentiments. Therefore, one of the main goals of the MYPLACE survey was

to analyse young people’s support for far right ideology: the prevalence of xenophobia,

welfare chauvinism, exclusionism, and overall negative attitudes towards minority groups.

In this chapter we look at how widespread, these attitudes are among youth, in our selected

localities, and what factors are driving them.

The national reports provide a mixed picture of the intra-country differences between our

research locations, as well as views of different socio-demographic groups. In some

countries one can observe large differences between the opinions of respondents in the two

research locations (e.g., Greece, Spain, Hungary, Latvia, Estonia, Russia, Finland, the UK),

whereas in others the differences are small or almost non-existent (e.g., Denmark).

Considering the similar cultural background of the localities within a country, one can

assume that certain contextual or compositional factors might explain the differences.

Nevertheless, there are some intriguing similarities that researchers have found in their

data, pointing to more general determinants of negative attitudes towards different ethnic

and religious minorities.

In some cases, the intra-country differences still do not obscure the distinctiveness of

attitudes of the young inhabitants of some countries. For example, citizens of post-Soviet

countries, Latvia, Estonia, and Russia, seem to have generally the most negative attitudes

towards minorities and immigrants, whereas young people in Germany hold the most

tolerant and inclusive opinions. The national reports also reveal a strong concern about the

growing immigration rates among young people in the South of Europe, especially in Greece

and Portugal, regardless of the location; resulting in calls for stricter border controls.

Moreover, xenophobia is also extremely widespread among youth in Russia.

While in Russia, one can talk about a general prejudice towards minorities and immigrants

as one attitudinal dimension, in other countries (e.g., Greece or Hungary) it is not

necessarily the case. National reports clearly show that all ethnic or religious minorities, are

not treated the same by the young people; some are more tolerated or respected than

others. In many of our surveyed countries (e.g., Denmark, Finland, Portugal, Spain, Russia,

Slovakia), the ethnic group most prone to discrimination are the Roma. In Greece, especially

the Argyroupoli area, the least trusted group are Albanians, who are the largest minority

group in the country. In post-Soviet, Estonia and Latvia, there is a large minority of Russians,

yet attitudes towards this group, despite the politicised issue of ethnic relations in these

MYPLACE August 2014


countries, are mostly positive. The report on Latvia highlights the role of history and media

in facilitating inter-ethnic tensions.

Several socio-demographic variables are consistently linked to negative attitudes towards

ethnic and religious minorities across the localities covered in the national reports. In many

countries (e.g., Finland, Germany, Russia, and the UK) individuals with low levels of

education and working class backgrounds are most likely to have negative attitudes towards

minorities and immigrants. Nevertheless, there are some exceptions: in Slovakia the lowest

education and socio-economic status group tends to have more positive views on the Roma,

which the authors explain by the fact that respondents of lower social classes, are more

likely to be Roma themselves. In several national reports (e.g., Latvia, Spain) researchers

note that attitudes towards immigrants and minorities are also linked to a person’s financial

situation, with those experiencing resource-stress bearing more negative views. In Russia,

researchers find that age or rather experience of meeting different ‘others’ can increase

tolerance and lessen prejudice. In addition, some reports (eastern Germany, Latvia) have

also highlighted the importance of political knowledge and interest in politics as factors

reducing prejudice towards immigrants and different ethnic and religious groups. In some

other countries (Greece, Russia, and the UK) male respondents were found to bear more

extreme and negative views towards minorities, than females.

Analysis of Southern European countries (Spain, Greece) indicates that attitudes towards

minorities can depend on the number of immigrants or minority residents in the area.

Nevertheless, for example in Denmark, the fact that there are more Muslims in Odense East

than Odense Centrum does not result in differing attitudes towards this group, and in Latvia

xenophobic attitudes are widespread, despite the very low immigration rates. Analysing the

Spanish localities, researchers note that the type of immigration might explain at least some

of the differences: most immigrants in Sant Cugat are from the richest European countries,

whereas Vic receives very a different kind of migrants, resulting in much higher support for

far right anti-immigrant parties in this location. The Russian national report suggests that

attitudes towards minorities and immigrants can also depend on the type and size of

locality: young people in Vyborg - a small, provincial and conservative locality – have more

anti-immigrant attitudes, whereas in Kupchino, which is a part of a large city, ethno-religious

tensions are more widespread.

Attitudes towards different ethnic minorities seem to vary by the country of origin or ethnic

group that the respondent himself belongs to, yet the effect is not consistent. In some cases

it is the majority ethnic group that holds comparatively more discriminatory attitudes

towards minorities (e.g. in Latvia, Slovakia, or the UK), in some – the non-majority groups

(e.g., in eastern Germany). A closer inspection of the data reveals that ethnic minorities

tend to have more positive attitudes towards their own group, whereas in general, ethnic

majority group seems to be more tolerant. A good example is Estonia, where Estonians are

MYPLACE August 2014


more accepting of the Roma, Jews, and Muslims, whereas other ethnic groups

(predominantly Russians) voice more positive views in respect to Russians. In Slovakia, too

researchers explain more positive attitudes towards Roma among minority respondents, by

the fact that many of them are Roma themselves. Similarly, in the UK, respondents in

Coventry, which is a much more ethnically diverse location, have greater tolerance towards

immigrants and more positive attitudes towards ethnic and religious minorities compared to

the more ethnically homogenous Nuneaton.

Finally, the national reports reveal that young people have generally positive views

regarding the contribution of immigrants to cultural diversity, and most often would support

granting them the same rights to welfare as the local population. Nevertheless, the issue of

jobs is by far the most sensitive. In many of the surveyed countries (e.g., Estonia, Greece,

Portugal, and Spain) despite overall positive attitudes towards immigrants, young people

believe that in a situation when jobs are scarce, employers should give priority to local

people.

Literature Negative attitudes towards minorities, xenophobia, as well as welfare chauvinism and

exclusionism, are characteristic features of far right ideology, often leading to social

exclusion, discrimination, or even open aggression and conflicts. Which social group is

disliked or stigmatised the most, differs from country to country. In Western Europe, much

attention has been drawn to immigrants from outside the European Union, especially

Muslims, in the context of the alleged ‘Islamisation’ of Europe or the debate on terrorism

and security (Mares, 2013; Yilmaz, 2012). East-Central countries have received criticism as

regards treatment of certain ethnic groups (e.g., Russians in Latvia and Estonia, Hungarians

in Romania and Slovakia, Turks in Bulgaria), yet the most victimised minority group in the

region is still the Roma (Mudde, 2005). For historical reasons, unlike in the western, in post-

communist countries the concept of ‘other’ revolves most around the ethnic groups that

settled in these countries during the Soviet times.

Not all countries have the same minority groups, and for historical reasons the

characteristics of these groups and attitudes towards them vary. In addition, the level of

development and welfare systems may influence the numbers and type of immigrants the

country attracts, as well as options for immigrants to integrate into the labour market. The

question of welfare provision to immigrants has recently been growing in importance in the

political agenda of Western European countries too.

Different explanations have been given to far right attitudes, some concentrating on

individual social background or socio-psychological characteristics of an individual, some -

socio-political or protest attitudes, or socio-structural explanations (immigration,

MYPLACE August 2014


unemployment etc.) and institutional factors (political opportunity structure). Previous

research has shown that those at the bottom of the social structure (working class, with a

lower level of education and income) due to their less secure position feel more insecure

and can be expected to bear more far-right attitudes. The dominant ethnic group, has been

shown in previous studies, to ascribe higher importance to national identity (Lewin-Epstein

and Levanon, 2005),. Whereas according to the ‘cultural affinity thesis’, those who are

themselves excluded from the dominant culture, are more likely to develop feelings of

acceptance and empathy towards members of other marginalised or minority groups

(Fetzer, 2000; Hewstone et al., 2002; Binder et al., 1997). Ethnic nationalism has also been

mentioned as one of the factors driving exclusionist attitudes, as well as overall

suspiciousness and lack of trust in people in general. Xenophobia and negative attitudes

towards minorities are often fuelled by populist ideology, and it thrives on ignorance and

irrational fear, as being most likely to appeal to people who have little knowledge and

understanding of politics, and little trust in politicians. According to the ‘contact hypothesis’

(Allport, 1954; Pettigrew, 1971; Fetzer, 2000), contact with minorities is supposed to lessen

extra-group hostility and prejudice. If there are more such people in the locality, there is

more information and contact with other ethnic groups, thus, less space for extra-group

stereotypes. Welfare state type can also have an effect on xenophobia and negative

attitudes towards minorities. Due to their historical legacies, post-soviet countries, can be

expected to be more xenophobic and prejudiced against minority groups. Also, low social

expenditure leads to economic insecurity (Mau et al., 2012) thus one can expect the liberal

regime type to facilitate xenophobia, more than the Scandinavian type that offers

comprehensive social insurance coverage. ‘Realistic group conflict’ theory (Sherif and Sherif

1953) holds that if people see poverty around them, the economy of their country going

downhill, and immigration rates increasing, resource stress will facilitate development of

xenophobia and negative attitudes towards minority groups.

Based on the literature, we include several individual and contextual variables in our

regression models predicting negative attitudes towards minorities, xenophobia, welfare

chauvinism and exclusionism. Among them are standard socio-economic variables such as

age; gender; education; income (how easy to cope on present household income) and

economic status (unemployed, student, other vs. employed). Considering that many young

people still have not entered the labour market, we use parents’ social class as a proxy for

young people’s social class. We also include in our analysis country of birth ([country],

other), citizenship ([country], other), and ethnic group (majority, other). Surprisingly, these

variables are not too closely related, which would make a simultaneous inclusion of all of

them in the model impossible.

Among attitudinal variables, ethnic nationalism was included in the model, coded as a

summary variable from how important it is, for being a citizen of [country], ’to have been

born in [country]’ and’have at least one ethnic [country] parent)’. Crombach Alpha values

MYPLACE August 2014


>0.7 in all countries except Estonia (0.62) confirmed that the items form a coherent and

reliable scale; Attitudes towards minority groups can also be influenced by overall trust in

fellow people, thus, we included the generalised trust question. To test the impact of

dissatisfaction with incumbent politicians, we calculated the political trust index, consisting

of trust in government, parliament, and parties. Besides ‘interest in politics’ we also include

in our model, political knowledge score, calculated as a summary index of how many of the

three knowledge questions (country-specific) the respondent answered correctly.

Considering the ‘contact hypothesis’, we also include in our model, contact with minorities

index (how many close friends are members of a different race / ethnic or minority group);

Another potentially important explanatory variable - media exposure – was coded as a

summary index of the intensity of following news in media.

Among the contextual variables, we include welfare state types (Kaariainen and Lehtonen,

2006). We also analyse the effect of immigration rates, changes in GPD during the recent

economic crisis, and local-level poverty rates.28 In addition, we include the percentage of

young people born outside the country (as a proxy for recent immigrant population) and

belonging to an ethnic minority.

12.1 Attitudes towards minorities (Q40)

Six questions were included in the MYPLACE survey that measure attitudes towards three

different minority groups - Roma, Jews, and Muslims – as well as immigrants in general,

using a 5-point scale from strongly agree to strongly disagree:

Q40_1 Roma, Gypsies and travellers make a positive contribution to society

(‘strongly agree, agree, neither agree nor disagree, disagree, strongly disagree’)

Q40_2 The police should be stricter with Roma / Gypsies / travellers

Q40_3 Jewish people talk too much about what happened to them during the

Holocaust

Q40_4 Jewish people make an important contribution to society

Q40_5 Muslims make a positive contribution to society

Q40_6 It is right to be suspicious of Muslims

Q41_3 Migrants greatly contribute to national cultural diversity

The items were recoded so that in all cases a higher value would indicate a more negative

attitude. Considering that the attitudes towards different minorities should not necessarily

correlate, an index was created from these seven questions, by summing the scores and

dividing it by the number of items.

28

The percentage of families who find it difficult or very difficult to get by on their present income, aggregated from the answers of respondents as this is the most precise available measure at the level of localities.

MYPLACE August 2014


A further analysis revealed that except for Georgia (0.31) and Hungary (0.54), Cronbach

Alpha values of the newly formed factor are above 0.6. An overall Conbach Alpha of 0.73

indicates a good internal consistency of the scale. Principal Component Analysis too

suggested a one-factor solution, based on Eigenvalues >1, which explains 44% of the

variation in the data. Spearman correlations (two-tailed) between the seven items are all

highly significant and range from 0.2 to 0.57. It means that if a person has a negative

attitude towards one minority group, he or she is more likely to have negative attitudes

towards other groups too.

MYPLACE August 2014


Q41: Negative attitudes towards minorities


The responses are normally distributed and provide a good scale from

0 to 4, where higher values indicate more negative attitudes towards

minorities.

Graph 12.1: Mean scores of negative attitudes towards minorities by

location

Narrative

The overall mean for negative attitudes towards minorities is 1.94

(n=12,615, sd=0.64). This varies by location, with the highest mean

score at Trnava (Slovakia) with 2.51, and the lowest mean scores at

Jena and Rostock (eastern Germany) with 1.21 and 1.36.

There are clear clusters of locations, with all locations in post-socialist

East-Central Europe showing above average scores. Attitudes towards

minorities are comparatively less negative in Forstate & Jaunbuve,

which is a very ethnically diverse location. Among Mediterranean

countries, only young people in Greek locations have quite negative

attitudes towards minorities, whereas almost all western locations are

clustered at the bottom of the ‘negative attitudes towards minorities’

scale.


notable exceptions were; Finland (with Lieksa and Nurmes and Kuopio

in Finland, 0.53 difference between two mean scores), Latvia

(Agenskalns and Forstate & Jaunbuve with 0.27 difference).

MYPLACE August 2014


There are issues with missing values in some countries, which is likely to be due to the

different representation of certain ethnic or religious groups in the country and, thus,

different contact and knowledge of these minorities.

The table below illustrates the proportion of missing cases by the three minorities and the

summary variable consisting of evaluations of all minorities and immigrants.

Table 12.1 Proportion of Missing Values by question

Country Roma Jews Muslims

All

minorities

Croatia 7.6 13.9 10.9 18.3

Denmark 6.3 11.0 2.7 14.2

Estonia 8.0 20.7 16.4 26.1

Finland 10.3 27.3 13.8 29.8

Georgia 12.9 41.0 22.2 45.7

Western Germany 23.9 15.0 6.5 30.1

Eastern Germany 18.2 8.9 4.7 21.6

Greece 7.9 27.4 14.6 31.2

Hungary 7.8 25.4 32.9 36.3

Latvia 3.3 7.3 7.0 11.2

Portugal 7.8 24.4 16.6 27.6

Russia 9.3 21.6 12.0 25.4

Slovakia 4.0 23.7 21.0 28.4

Spain 9.7 20.9 6.2 24.1

UK 3.9 8.3 3.0 10.4

Total 9.3 19.9 13 22.5

The number of missing values is the highest in Georgian, Finnish, western German,

Hungarian, Slovak, Greek and Portuguese locations: more than 25 % could not give an

answer as regards to at least one of the minorities.

If we look separately at evaluations of different minorities, we see that young people in

Germany have found it difficult to formulate an opinion on the Roma; Georgians, Finns and

Greeks very often had no opinion on Jews; while Hungarians struggled to answer questions

on Muslims.

12.2 Attitudes towards Roma, Jews and Muslims

Considering the variation between countries in the size of certain minority groups, we also

calculated separately three factors measuring:

1. Negative attitudes towards Roma,

MYPLACE August 2014


2. Negative attitudes towards Jews and

3. Negative attitudes towards Muslims.

To learn about respondents’ attitudes towards Roma, they were asked, using a 5-point scale

from ‘strongly agree’ to ‘strongly disagree’, if:

Q40_1 Roma, Gypsies and travellers make a positive contribution to society

(‘strongly agree, agree, neither agree nor disagree, disagree, strongly disagree’)

Q40_2 The police should be stricter with Roma / Gypsies / travellers

Attitudes towards Jews were measured on the basis of following items:

Q40_3 Jewish people talk too much about what happened to them during the

Holocaust

Q40_4 Jewish people make an important contribution to society

And, finally, attitudes towards Muslims – with the following items:

Q40_5 Muslims make a positive contribution to society

Q40_6 It is right to be suspicious of Muslims

The items were recoded so that in all cases a higher value would indicate a more negative

attitude. An index was created from these two questions, by summing the scores and

dividing it by two.

The two items measuring attitudes towards Roma and Muslims were highly correlated, and

can be said to form an acceptable scale (Cronbach Alpha>0.6), whereas the two items

measuring attitudes towards Jews Q40_3 and Q40_4 are not so closely related (Spearman’s

correlation 0.25, Cronbach Alpha 0.41). Nevertheless, they both indicate negative

associations with Jews, so we code them together.

MYPLACE August 2014


Q40_1 & Q40_2: Attitudes towards the Roma


A Spearman rank correlation of 0.45 suggests that there is a

strong relationship between the two variables (Q40_1 and

Q40_2) and they can be combined. The scale ranges from 0 (least

negative attitudes towards Roma) to 4 (most negative attitudes

towards Roma).

Graph 12.2a: Mean scores of negative attitudes towards Roma

Narrative

The overall mean for negative attitudes towards the Roma, is

2.32 (n=15,366, sd=0.89). This varies by location, with the

highest mean score in Trnava (Slovakia) with 3.21 out of possible

4, and the lowest mean score in Jena (eastern Germany) with

2.47, which is still more negative than positive.

In regards to attitudes towards the Roma, there are no distinct

clusters of locations; Overall, young people in locations in Spain

and Germany have the least negative attitudes towards the

Roma, whereas those living in Hungary, Slovakia, Greece and

Russia that have a considerable Roma minority, as well as those

living in Lieksa and Nurmes (Finland) have the most negative

attitudes.


locations; notable exceptions were; Finland (Kuopio and Lieksa

and Nurmes), 0.49 difference between two mean scores, and

Slovakia (Trnava and Rimavska Sobota), 0.53 difference.

MYPLACE August 2014


Q40_3 & Q40_4: Attitudes towards Jews


A Spearman rank correlation of 0.25 suggests that the two

variables (Q40_3 and Q40_4) are only weakly related, yet based

on the fact that they both indicate negative associations with

Jews, they can still be combined. The scale ranges from 0 (least

negative attitudes towards Jews) to 4 (most negative attitudes

towards Jews).

Graph 12.2b: Mean scores of negative attitudes towards Jews

Narrative

The overall mean for negative attitudes towards Jews, is 1.84

(n=13,562, sd=0.86), while 20% were excluded from the analysis

because they could not answer at least one of the two questions.

Attitude towards Jews varies by location, with the highest mean

score in Ozd (Hungary) with 2.32 out of possible 4, and the

lowest mean score in Rostock and Jena (eastern Germany) with

1.35 and 1.24 accordingly.

In regards to attitudes towards Jews, there are distinct clusters

of locations; young people in locations in Western Europe,

especially Germany, voice negative attitudes towards Jews less

often.

The mean scores between the two locations differ somewhat in

almost all countries, but not by much.

MYPLACE August 2014


Q40_5 & Q40_6: Attitudes towards Muslims Graph 12.2c: Mean scores of negative attitudes towards Muslims


A Spearman rank correlation of 0.42 suggests that there is a moderately

strong relationship between the two variables (Q40_4 and Q40_5) and they

can be combined. The scale ranges from 0 (least negative attitudes towards

Muslims) to 4 (most negative attitudes towards Muslims).

Narrative

The overall mean for negative attitudes towards Muslims, is 1.87 (n=14,725,

sd=0.87) out of possible 4, thus, it is more positive than negative. This varies

by location, with the highest mean score in Vyborg (Russia) with 2.56, and

the lowest in Rostock and Jena (eastern Germany) with 1.26 and 1.17.

There are clear clusters of locations; young people in all locations in post-

socialist countries in East-Central Europe, as well as those from Greece, the

attitudes towards Muslims are more negative than on average, whereas

those living in Western Europe, in particular Germany, the UK and Denmark,

have the least negative attitudes.


exceptions were; Latvia (Agenskalns and Forstate & Jaunbuve, 0.39

difference between two mean scores), and Finland (Lieksa and Nurmes and

Kuopio, 0.64 difference).

MYPLACE August 2014


12.3 Xenophobia (Q41)

Respondents were asked, using a 5-point scale from ‘strongly agree’ to ‘strongly disagree’, if

[country] should have stricter border controls and visa restrictions to prevent further

immigration’.

MYPLACE August 2014


Q41_1: Xenophobia (Q41_1) Graph 12.3: “[Country] should have stricter border controls and visa

restrictions to prevent further immigration by location



responses to the question ‘[Country] should have stricter border

controls and visa restrictions to prevent further immigration’.

Narrative


respondents who ‘strongly agree’, 32% ‘agree’, 22% ‘neither agree nor

disagree’, 22% ‘disagree’ and 6% ‘strongly disagree’, that [Country]

should have stricter border controls and visa restrictions to prevent

further immigration.

Young people in the Greek locations have the most xenophobic

attitudes, with almost 90% supporting limitation of further

immigration. They are followed by Vyborg (Russia), Bareiro (Portugal),

Trnava (Slovakia), and Nuneaton (UK), whereas German locations are

clustered in the bottom of the xenophobia scale.

There are no clear clusters of locations by country, as the localities

within one country vary quite a lot, especially, the UK (Nuneaton and

Coventry), Finland (Lieksa and Nurmes and Kuopio), as well as Latvian,

Hungarian and Slovakian locations. As a result, there are also some

Western European locations that are characterised by xenophobia

(Nuneaton, UK), while some of the least xenophobic are locations in

the East of Europe (Forstate & Jaunbuve, Latvia; Narva, Estonia).

MYPLACE August 2014


12.4 Exclusionism (Q41_2, Q41_4, Q41_5)

Exclusionism was operationalised on the basis of three items that describe exclusion from

three different resources: jobs, land, and welfare. They were measured on a 5-point scale

from strongly agree to strongly disagree:

Q41_2 Foreigners should not be allowed to buy land in [COUNTRY] (‘strongly agree,

agree, neither agree nor disagree, disagree, strongly disagree’)

Q41_4 Migrants should have the same rights to welfare (health care, housing,

education) as people from [COUNTRY]

Q26_5 When jobs are scarce, employers should give priority to [COUNTRY] people

over foreign workers

The items were recoded so that in all cases a higher value would indicate a more

exclusionist attitude. Considering that the exclusionist attitudes as regards to different

resources should not necessarily correlate, an index was created from these three

questions, by summing the scores and dividing it by the number of items.

MYPLACE August 2014


Q41_1a: Exclusionism


The new factor scale has a minimum of zero (and a maximum of

4.

Graph 12.4: Mean score for exclusionism

Narrative

The overall mean for exclusionism is 1.95 (n=15,978, sd=0.90) out

of a possible 4. This varies by location, with the highest mean

score in Ozd (Hungary) with 2.83, and the lowest mean score in

Jena (eastern Germany) with 0.99.

There are clear clusters of locations; with locations Hungary at the

top of the exclusion scale, followed by other post-soviet localities

in Russia, Georgia, and Slovakia. Locations in Germany are

clustered in the bottom, young people being the least exclusionist

towards immigrants.


locations; notable exceptions were; Finland (Lieksa and Nurmes

and Kuopio, 0.81 difference between two mean scores), and the

UK (Nuneaton and Coventry, 0.47 difference between two mean

scores).

MYPLACE August 2014


12.5 Welfare chauvinism (Q41_2)

Welfare chauvinism was measured by a question ’Migrants should have the same rights to

welfare (health care, housing, education) as people from [COUNTRY]’ which was reversed,

for more positive answers to mean more welfare chauvinistic attitudes.

MYPLACE August 2014


Q41_2: Welfare chauvinism (Q41_2)



responses to the question ’Migrants should NOT have the same rights to

welfare (health care, housing, education) as people from [COUNTRY]’.

Graph 12.5: Migrants should NOT have the same rights to welfare

(health care, housing, education) as people from [COUNTRY]by location

Narrative



disagree’, 41% ‘disagree’ and 14% ‘strongly disagree’, that migrants should

NOT have the same rights to welfare (health care, housing, education) as

people from [COUNTRY].

Vyborg (Russia) had the largest proportion of respondents (48%) voicing

welfare chauvinist attitudes, followed by Kupchino (Russia, 47%), and

Georgian and Hungarian locations. Koupio (Finland) had the smallest

proportion (7%) of people agreeing that migrants should NOT have the same

rights to welfare (health care, housing, education) as people from Finland.

German and Danish localities cluster at the bottom of the welfare

chauvinism scale, but they are joined also by Koupio (Finland), Vic (Spain)

and Narva (Estonia).

There are no clear clusters of locations by country, as the localities within

one country vary quite a lot, for example there is a large difference in

welfare exclusionist attitudes between Nuneaton and Coventry (UK), Narva

and Tartu (Estonia), Kuopio and Lieksa and Nurmes (Finland).

MYPLACE August 2014


12.6. Multi-level Modelling

Model A: Negative attitudes towards minorities (Q40_1…Q40_6; Q41_3)

Due to the high number of missing cases in some of the questions, 33% of the cases were

excluded from the regression analysis. The initial analysis of variance components reveals

that 71% of variance in negative attitudes towards minorities can be explained by

differences between individuals, whereas 29 percent of the variance lies at the level of

localities (intra-class correlation (ICC) = 0.29) (Table 12.6.1). The substantial variation at the

level of localities means that it is necessary to search for contextual variables that would

help to explain it.

Inclusion of the previously described individual level variables resulted in a 12 percent

decrease of residual, individual level variance, and led to an even larger – 50 percent

reduction in intercept variance. It means that much of the observed differences between

localities are due to the composition effect, i.e., differing individual characteristics of their

young residents. In the next step, we added a number of contextual variables, which helped

to achieve a massive 68 percent reduction of intercept variance compared to the model

containing only the individual level predictors – again, an improvement that is highly

significant. This means that our selected contextual variables are capable of explaining a

huge portion of the unexplained variation in negative attitudes towards minorities between

localities. Only 7% of unexplained variance remains at the level of localities, although it is

still significant (at 0.01 level).

Table 12.6.1. Variance components of regression models explaining negative attitudes

towards minorities Empty

model

(random

intercept

only)

With

individual

level

explanatory

var’s

With

country

level

explanatory

var’s


Negative attitudes towards

minorities

σ (country level) 0.118 0.058 0.019

-loglikelihood 18536.315 16802.549 16852.004

Turning to the substantive results (Table 12.6.2), as expected, we find that ethnic

nationalism is among the strongest predictors of negative attitudes towards minorities. A

one-unit increase in ethnic nationalism on a 5-point scale is associated with a 0.133 unit

increase in xenophobia, measured on the same scale. One can conclude that ethnic

MYPLACE August 2014


nationalism is not just exclusionist, with regards to those who do not have ethnic or national

roots, but it is also linked to overall negative attitudes towards minorities.

Men much more than women are likely to have negative attitudes towards minorities, as

are people in less advantageous social-economic positions, whose family is struggling

financially and whose parents belong to upper or lower middle class. Such attitudes are also

linked to lower levels of education: young people with tertiary education are less likely to

have negative attitudes towards minorities. Interestingly though negative attitudes towards

minorities are more widespread among those who are employed, which could be linked to

problems of communication and cultural or language barriers with the minorities. At the

same time, our results do not confirm the assumptions of psychological theories regarding

the psychological susceptibility of youth to populist far right ideology. Age is not a significant

predictor of negative attitudes towards minorities among youth.

Table 12.6.2. Determinants of negative attitudes towards minorities, n2=30, n1=11454 Model A

b SE


Gender (a man) .074 *** (.010)

Age .002 (.002)

Tertiary education Ref. (Ref.

Upper secondary education .088 *** . ( 018)

Less than upper secondary education .049 *** (.014)

Employed Ref. Ref.

Unemployed -.030 (.023)

In education -.033 ~ (.019)

Other -.065 *** (.013)

Class 3 Ref. Ref.

Class 2 .062 *** (.014)

Class 1 .075 *** (.018)

Class 0 .028 (.018)

Financial problems in the household .029 * (.012)

Dominant ethnicity in the country of residence .037 * (.018)

Citizen of the county of residence .015 (.026)

Born in the county of residence -.129 *** (.025)

Ethnic nationalism .133 *** (.005)

Political knowledge -.024 *** (.006)

Interest in politics -.042 *** (.006)

Media exposure .014 *** (.002)

Social trust .-012 *** (.002)

Political trust -.030 *** (.002)

Three or more friends of other ethnicity/race Ref. Ref.

Two friends of other ethnicity/race .143 *** (.013)

One friend of other ethnicity/race .067 *** (.016)

No friends of other ethnicity/race .080 *** (.016)

Contextual predictors: Ref. Ref.

GDP growth during the crisis (2008-2011) .002 (.004)

% of families in the area experiencing financial difficulties .007 ** (.002)

% of young people born abroad -.002 (.004)

MYPLACE August 2014


% of young people of non-majority ethnicity -.002 (.002)

Net immigration rates .006 (.017)

Size of the area (thousands) .004 (.003)

Liberal Ref. Ref.

Post-socialist .165 (.154)

Nordic .101 (.151)

Conservative -.176 (.146)

Mediterranean -.203 (.144)

Constant 1.365 *** (.176)


In line with the ‘cultural affinity thesis’, the analysis confirms that negative attitudes

towards minorities are more common among the dominant ethnic group than among those

who belong to minorities themselves, regardless of citizenship status. However, controlled

for ethnic group, those born outside the country tend to have more negative attitudes

towards minorities. This relationship is intriguing, and there could be several explanations to

it, such as cultural differences or perceived inter-group conflict or competition.

The analysis confirms the role of information in lessening prejudice towards minority

groups: those who are more interested in, and better informed about, politics are less likely

to hold negative attitudes towards minorities. Among the strongest predictors of anti-

minority attitudes are also political distrust and, interestingly, media exposure, with those

who watch different media channels more frequently having more negative attitudes

towards minorities. It shows that media are at least partly responsible for the recent spread

of populism and anti-minority sentiments across Europe.

Attitudes towards all kinds of minorities are more tolerant if people have more contact with

minorities themselves. In line with the ‘contact hypothesis’, we find that those who have

friends of another race or ethnicity are much less likely to have negative attitudes towards

minorities. However, one must acknowledge that there is a problem of endo-geneity. The

model also does not allow us to disentangle the causal relationships of the two variables – it

is possible than people have more friends of other ethnicity, because they have more

positive attitudes towards minorities. We also find that people who distrust others in

general are more likely to have negative attitudes towards minorities – they might feel that

minorities are ‘different’, more unlike themselves, and thus they can be trusted less. If

overall trust in people will increase, the attitudes towards minorities will improve too.

We also find strong support for the ‘realistic group conflict’ theory. As mentioned earlier,

those who are poor tend to have more negative attitudes towards minorities, and the

overall situation in the area seems to matter too. The number of families in the area facing

financial difficulties is the strongest contextual level predictor of negative attitudes towards

minorities, and the only one that is statistically significant. These attitudes are likely to be

driven by resource stress and competition. Other contextual level factors such as the

welfare state type are not significant predictors of negative attitudes towards minorities.

MYPLACE August 2014


Model B: Negative attitudes towards Roma, Jews and Muslims separately (Q40_1, 2; Q40_3,

4; Q40_5, 6)

In this chapter, we look at determinants of negative attitudes towards different minority

groups – Roma, Jews and Muslims – separately.

First, the analysis of variance components reveals that while attitudes towards Jews can

largely be explained by differences between individuals, attitudes towards Muslims are to a

significant extent influenced by context: accordingly, 14% and 23% of variance in negative

attitudes towards these groups lies at the level of localities (Table 12.6.3).

Inclusion of the previously described individual level variables resulted in an 5 (Roma) to 8

(Muslims) percent decrease of residual, individual level variance, meaning that many more

individual factors than those included in the model affect these perceptions. Although the

reduction might seem small, the chi-square test of deviances, confirms that it is statistically

significant. Interestingly, adding individual level variables led to an even larger – 35 (Roma)

to about 50 percent (Muslims, Jews) reduction in intercept variance. It means that much of

the observed differences between localities are due to the composition effect, i.e., differing

individual characteristics of their young residents. In the next step, we added a number of

contextual variables, which helped to achieve a massive 74 percent reduction of intercept

variance compared to the model containing only the individual level predictors as regards to

attitudes towards Jews, 42 percent – Muslims, and only 35 percent – Roma. This means that

our selected contextual variables are capable of explaining a huge portion of the

unexplained variation in negative attitudes towards Jews between localities, yet there are

more important contextual factors responsible towards negative attitude towards Roma

and Muslims that are not included in the model. Only 2% of unexplained variance in

attitudes towards Jews remains at the level of localities, although one must note that it is

still significant (at 0.01 level).

Table 12.6.3. Variance components of regression models explaining negative attitudes

towards Roma, Jews and Muslims Empty

model

(random

intercept

only)

With

individual

level

explanatory

var’s

With

country

level

explanatory

var’s

σ (individual level) .648 .614 .614


Roma

σ (country level) .130 .085 .055

-loglikelihood 32807.97 31427.81 31486.5



Jews


MYPLACE August 2014


-loglikelihood 26187.69 24785.37 24834.11



Muslims


-loglikelihood 30351.72 28474.59 28523.76

Turning to the substantive results (Table 12.6.4), we find that ethnic nationalism is among

the strongest predictors of negative attitudes towards all kinds of minorities among youth,

especially towards Muslims. A one-unit increase in ethnic nationalism on a 5-point scale is

associated with a 0.152 unit increase in negative attitudes towards Muslims, measured on

the same scale. The results also show that men and those who are employed, are more

likely to have negative attitudes towards all kinds of minorities.

Low level of income and education is linked with negative attitudes towards Muslims and

Jews, but the differences are less pronounced as regards to the Roma. Similarly, upper and

lower middle class is more likely to hold negative attitudes towards Jews and Muslims, while

with regards to the Roma this is mostly the lower middle class. Interestingly though those

who are employed are more xenophobic than students or unemployed.

MYPLACE August 2014


Table 12.6.4. Determinants of negative attitudes towards Roma, Jews and Muslims, n2=30

Model B1

Roma

n1=13590

Model B2

Jews

n1=12205

Model B3

Muslims

n1=11879

b SE b SE b SE


Gender (a man) .087 *** .014 .090 *** .013 .037 ** .013

Age -.001 .003 .001 .003 .006 ~ .003

Tertiary education Ref. Ref. Ref. Ref. Ref. Ref.

Upper secondary education -.005 .026 .086 *** .024 .126 ** .025

Less than upper secondary education .036 ~ .021 .022 .019 .044 * .019

Employed Ref. Ref. Ref.

Unemployed -.012 .032 .026 .029 -.063 * .030

In education -.072 ** .026 -.011 .024 -.023 .025

Other -.049 ** .019 -.064 *** .017 -.076 *** .018

Class 3 Ref. Ref. Ref. Ref. Ref. Ref.

Class 2 .003 .019 .093 *** .018 .081 *** .018

Class 1 .062 * .025 .061 ** .023 .087 *** .024

Class 0 .033 .025 -.003 .023 .030 .024

Financial problems in the household -.023 .017 .040 * .016 .064 *** .017

Dominant ethnicity in the country of residence .162 *** .026 -.085 *** .024 .036 .025

Citizen of the county of residence .008 .038 -.048 .034 .070 * .035

Born in the county of residence -.123 *** .036 -.237 *** .032 -.123 *** .033

Ethnic nationalism .132 *** .007 .101 *** .006 .152 *** .007

Political knowledge .024 ** .008 -.044 *** .008 -.046 *** .008

Interest in politics -.046 *** .009 -.032 *** .008 -.042 *** .009

Media exposure .014 *** .003 .010 *** .003 .017 *** .003

Social trust -.013 *** .003 -.008 ** .003 -.015 *** .003

Political trust -.030 *** .004 -.026 *** .003 -.032 *** .003

Three or more friends of other ethnicity/race Ref. Ref. Ref. Ref. Ref. Ref.

Two friends of other ethnicity/race .183 *** .019 .075 *** .017 .216 *** .018

One friend of other ethnicity/race .086 *** .023 .035 ~ .021 .121 *** .022

No friends of other ethnicity/race .089 *** .024 .069 ** .021 .100 *** .022

Contextual predictors: Ref. Ref. Ref.

GDP growth during the crisis (2008-2011) .006 .007 .000 .003 .000 .005

% of families in the area experiencing

financial difficulties .008 * .004 .005 * .002 .007 * .003

% of young people born abroad -.005 .008 .002 .003 -.002 .006

% of young people of non-majority ethnicity .000 .003 -.004 ** .001 -.004 .003

Net immigration rates .037 .030 .006 .013 -.014 .023

Size of the area (thousands) .005 .005 .000 .002 .004 .004

Liberal Ref. . Ref. Ref.

Post-socialist -.167 .263 .305 * .112 .348 .202

Nordic .001 .258 .073 .110 .213 .198

Conservative -.490 ~ .251 -.003 .107 .015 .192

Mediterranean -.533 * .247 -.003 .105 -.003 .189

Constant 1.893 *** .294 1.646 *** .146 .977 *** .231


MYPLACE August 2014


As regards the ‘cultural affinity thesis’, the analysis shows that majority ethnic group is more

likely to have negative attitudes towards the Roma, and less negative - towards Jews,

whereas citizenship status mostly worsens attitudes towards Muslims. Considering that

most Roma and Jews are residing in the country for a long time, one can conclude that in

the latter case the negativism is directed towards recent immigrants –extra-group

representatives who do not have country citizenship. In all cases, those born outside the

country bear more negative attitudes towards all kinds of minority groups. It is possible that

they try to distance themselves from them, yet there are also other psychological

mechanisms that could be responsible for that.

Those who are more interested in, and better informed about, politics, and watch different

media channels less frequently are less likely to hold negative attitudes towards all kinds of

minorities. Negative attitudes towards all minorities are also strongly linked to social and

political distrust, showing that overall suspiciousness and mistrust facilitates prejudice and

hatred of other ethnic groups.

Attitudes towards all kinds of minorities are more tolerant if people have more contact with

minorities themselves. In line with the ‘contact hypothesis’, we find that those who have

friends of another race or ethnicity are much less likely to have negative attitudes towards

any minorities, although, as mentioned before, it is possible that people who have more

friends of other ethnicities, do so because they have more positive attitudes. Young people

in localities where there are more people of ethnic minorities around tend to have less

negative attitudes towards Jews, yet it has no effect on attitudes towards the Roma or

Muslims.

Again, we find support for the ‘realistic group conflict’ theory in regards to all minorities: the

number of families in the area facing financial difficulties is among the strongest contextual

level predictors of negative attitudes.

Attitudes towards different minorities vary depending on the type of countries. Attitudes

towards the Roma are much more negative in the liberal regime type (the UK) than in

conservative (Germany) or Mediterranean countries. Attitudes towards Jews and Muslims

are the most negative among young people in post-socialist countries (although in the latter

case the difference from liberal regimes, i.e., the reference category is not statistically

significant). It confirms that cultural context significantly affects which groups will be

discriminated against.

Model C: Xenophobia (Q41_1)

Our dependent variable is ‘xenophobia’, measured in the MYPLACE survey by ‘[country]

should have stricter border controls and visa restrictions to prevent further immigration’.

MYPLACE August 2014


The initial analysis of variance components reveals that 79 percent of variance in

xenophobia scores can be explained by differences between individuals, whereas 21 per-

cent of the variance lies at the level of localities (intra-class correlation (ICC) = 0.21) (Table

12.6.5). The substantial variation at the level of localities means that it is necessary to

employ a multilevel analysis to try to explain it.

Inclusion of the previously described individual level variables resulted in an 8 percent

decrease of residual, individual level variance. Although the reduction might seem small, the

chi-square test of deviances (df = 25, loglikelihood = 2367) confirms that it is statistically

significant. Interestingly, adding individual level variables, led to an even larger – 41 percent





containing only the individual level predictors. This means that our selected contextual

variables are capable of explaining a huge portion of the unexplained variation in

xenophobia scores between localities. Only 5% of unexplained variance remains at the level

of localities, although one must note that it is still significant (at 0.01 level).

Table 12.6.5. Variance components of regression models explaining xenophobia Empty

model

(random

intercept

only)

With

individual

level

explanatory

var’s

With

country

level

explanatory

var’s


Xenophobia σ (country level) 0.279 0.164 0.051

-loglikelihood 39489.417 39442.544 39489.417

Turning to the substantive results (Table 12.6.6), in line with our hypothesis, we find that

ethnic nationalism is among the strongest predictors of xenophobia among youth. A one-

unit increase in ethnic nationalism on a 5-point scale is associated with a 0.235 unit increase

in xenophobia, measured on the same scale. One can conclude that ethnic nationalism is

not just exclusionist, with regards to those who do not have ethnic or national roots, but it is

also linked to anti-immigration sentiments.

Similar to other studies, we find that men are more likely to be xenophobic than women, as

are people in less advantageous social-economic positions, i.e., young people with a low

level of education, whose family is struggling financially and whose parents do not belong to

the highest social class. Interestingly though those who are employed are more xenophobic

than students or unemployed. It could be that they fear that immigration would threaten

MYPLACE August 2014


their position in the labour market. At the same time, our results do not confirm the

assumptions of psychological theories regarding the psychological susceptibility of youth to

populist far right ideology. Age is not a significant predictor of xenophobia among youth,

i.e., adolescents are as likely to hold xenophobic views as those who are older.

As regards the ‘cultural affinity thesis’, the analysis confirms the hypothesis only partly.

Xenophobia is, indeed, linked to citizenship status, which is not surprising considering that

those who are citizens can be said to have most to fear and to protect in terms of status.

However, xenophobia does not depend on whether the person is a member of a majority or

minority group, as well as on his or her country of birth (Sig.<0.05). There is no extra

empathy from the minority ethnic groups towards potential future immigrants.

In line with the ‘populist thesis’, the analysis confirms our expectations: those people more

interested in and better informed about politics are less likely to hold xenophobic attitudes.

Xenophobia is also strongly linked to political distrust and, interestingly, media exposure is

an important predictor of anti-immigrant attitudes: those who watch different media

channels more frequently, are more likely to hold xenophobic attitudes. It shows that media

are at least partly responsible for the recent spread of xenophobia across Europe.

MYPLACE August 2014


Table 12.6.6. Determinants of xenophobia; n2=30, n1=14419 Model C

b SE


Gender (a man) 0.063 *** (0.017)

Age -0.001 (0.004)

Tertiary education Ref. Ref.

Upper secondary education 0.048 ~ (0.026)

Less than upper secondary education 0.098 ** (0.032)

Employed Ref. Ref.

Unemployed -0.067 * (0.032)

In education -0.086 *** (0.023)

Other 0.007 (0.039)

Class 3 Ref. Ref.

Class 2 0.062 * (0.031)

Class 1 0.129 *** (0.031)

Class 0 0.109 *** (0.024)

Financial problems in the household 0.083 *** (0.022)

Dominant ethnicity in the country of residence 0.046 (0.032)

Citizen of the county of residence 0.203 *** (0.047)

Born in the county of residence 0.061 (0.044)

Ethnic nationalism 0.235 *** (0.009)

Political knowledge -0.023 * (0.010)

Interest in politics -0.046 *** (0.011)

Media exposure 0.011 ** (0.004)

Social trust -0.007 ~ (0.004)

Political trust -0.021 *** (0.004)


Two friends of other ethnicity/race 0.057 ~ (0.029)

One friend of other ethnicity/race 0.063 * (0.028)

No friends of other ethnicity/race 0.153 *** (0.023)


GDP growth during the crisis (2008-2011) 0.006 (0.006)

% of families in the area experiencing financial difficulties 0.015 *** (0.004)

% of young people born abroad -0.002 (0.007)

% of young people of non-majority ethnicity -0.006 ~ (0.003)

Net immigration rates 0.080 * (0.029)

Size of the area (thousands) 0.001 * (0.001)

Liberal Ref. Ref.

Post-socialist -0.629 * (0.253)

Nordic -0.622 * (0.248)

Conservative -0.644 * (0.241)

Mediterranean -0.526 * (0.237)

Constant 1.480 *** (0.291)


Attitudes towards all kinds of minorities are more tolerant, and xenophobia decreases, if

people have more contact with minorities themselves. In line with the ‘contact hypothesis’,

we find that those who have three or more friends of another race or ethnicity are much

less likely to be xenophobic. However, one must acknowledge that there is a problem of

endo-geneity. In addition, the model does not allow us to disentangle the causal

MYPLACE August 2014


relationships of the two variables – it is possible than people have more friends of other

ethnicity because they are less anti-immigrant. However, the ‘contact hypothesis’ is also

supported by the fact that xenophobia is lower in localities where there are more people of

ethnic minorities around (although the significance slightly misses the conventional 0.5

level).

We also find strong support for the ‘realistic group conflict’ theory. Not only are those who

are poor more xenophobic, but the overall situation in the area seems to matter too. The

number of families in the area facing financial difficulties is among the strongest contextual

level predictors of xenophobia: if the number of poor families in the area increased by 20

percent it would mean a 0.3 units increase in xenophobia. At the same time, the effect of

GDP growth during the crisis is insignificant. It shows that resource stress is mainly

experienced because of the economic conditions young people see around them, not

general macro-economic developments.

As expected, xenophobia is higher where the net immigration rates are higher, especially in

the bigger cities where immigrants often settle. At the same time, the analysis does not

confirm that post-socialist status would be linked to increased xenophobia. This means that

other factors, rather than post-socialist status, are responsible for high rates of xenophobia

in post-communist countries in East-Central Europe, such as resource stress, low quality of

media, and little social and political trust, combined with little contact with people from

other ethnic groups. In line with our expectations, we find that the liberal welfare state

type, such as observed in the UK, is more likely to produce xenophobia than other types of

welfare regimes. A potential explanation might be that people in less liberal countries that

provide more social security feel more secure and less threatened by the effect of

immigration on their life chances. From this point-of-view, the dismantling of the European

Social Model observed during the recent economic crisis risks facilitating the spread of anti-

immigrant attitudes across European youth.

Model D: Exclusionism (Q41_2, Q41_4, Q41_5)

The initial analysis of variance components reveals that 66 percent of variance in

exclusionism scores can be explained by differences between individuals, whereas 34

percent of the variance lies at the level of localities (intra-class correlation (ICC) = 0.34)

(Table 12.6.7). The substantial variation at the level of localities means that it is necessary to

employ a multilevel analysis to try to explain it.


decrease of residual, individual level variance. The chi-square test of deviances (df = 25,

loglikelihood = 2454) confirms that it is statistically significant. Interestingly, adding

individual level variables led to an even larger – 50 percent reduction in intercept variance.

MYPLACE August 2014


It means that much of the observed differences between localities are due to the

composition effect, i.e., differing individual characteristics of their young residents. In the

next step, we added a number of contextual variables, which helped to achieve a massive 84

percent reduction of intercept variance compared to the model containing only the

individual level predictors. This means that our selected contextual variables are capable of

explaining a huge portion of the unexplained variation in xenophobia scores between

localities. Only 4% of unexplained variance remains at the level of localities, although one

must note that it is still significant (at 0.01 level).

Table 12.6.7. Variance components of regression models explaining exclusionism Empty

model

(random

intercept

only)

With

individual

level

explanatory

var’s

With

country

level

explanatory

var’s


Exclusionism σ (country level) .278 .138 .022

-loglikelihood 31571.57 29117.87 29146.22

Ethnic nationalism is among the strongest predictors of exclusionism among youth (Table

12.6.8). A one-unit increase in ethnic nationalism on a 5-point scale is associated with a 0.2

unit increase in exclusionism, measured on the same scale. One can conclude that ethnic

nationalism is linked to exclusionist views.

As in case of other variables, men are more likely to be exclusionist than women, as are

young people with less than a tertiarylevel education, whose family is struggling financially.

Interestingly though those who are employed or students are more exclusionist than those

who are unemployed or economically inactive. It could be that they are trying to protect

their advantageous position in the labour market. Contrary to the assumptions of

psychological theories, we find that age is positively related to exclusionist attitudes i.e.,

adolescents are less likely to hold exclusionist views as those who are older. Such attitudes

might be linked to a sense of growing competitiveness for resources.

The analysis confirms the ‘cultural affinity thesis’. Exclusionism as regards to immigrants is

indeed less widespread among those who are not the majority ethnicity, are born outside

country, and are not citizens themselves. On the other hand, this can be interpreted as

concern about one’s own situation and willingness not to be excluded.

Young people who are interested in politics, and watch media less frequently, are less likely

to hold exclusionist attitudes. Exclusionism is also strongly linked to social and political

distrust, with those who have little trust in politicians and people in general being more

MYPLACE August 2014


exclusionist. Perhaps they fear that immigrants could abuse the system in a dishonest

manner.

Like overall negative attitudes towards minorities and immigrants, exclusionism decreases if

people have more contact with minorities themselves. In line with the ‘contact hypothesis’,

we find that those who have friends of another race or ethnicity are much less likely to be

exclusionist. The ‘contact hypothesis’ is also supported by the fact that exclusionism is lower

in localities where there are more people around who were born abroad.

MYPLACE August 2014


Table 12.6.8. Determinants of exclusionism; n2=30, n1=14165 Model D

b SE


Gender (a man) .043 *** .012

Age .008 ** .003


Upper secondary education .133 *** .022

Less than upper secondary education .079 *** .018

Employed Ref. Ref.

Unemployed -.059 * .027

In education .016 .022

Other -.079 *** .016

Class 3 Ref. Ref.

Class 2 .046 ** .017

Class 1 .031 .022

Class 0 .018 .022

Financial problems in the household .051 ** .015

Dominant ethnicity in the country of residence .074 *** .022

Citizen of the county of residence .165 *** .032

Born in the county of residence .117 *** .030

Ethnic nationalism .200 *** .006

Political knowledge -.007 .007

Interest in politics -.033 *** .008

Media exposure .008 ** .003

Social trust -.019 *** .003

Political trust -.026 *** .003


Two friends of other ethnicity/race .179 *** .016

One friend of other ethnicity/race .116 *** .020

No friends of other ethnicity/race .065 ** .020


GDP growth during the crisis (2008-2011) .006 .004

% of families in the area experiencing financial difficulties .003 .002

% of young people born abroad -.011 * .005

% of young people of non-majority ethnicity .004 ~ .002

Net immigration rates .031 .019

Size of the area (thousands) .003 .003

Liberal Ref. Ref.

Post-socialist .254 .170

Nordic -.240 .166

Conservative -.560 ** .162

Mediterranean -.218 .159

Constant .699 ** .197


As regards to exclusionism, we find limited support for the ‘realistic group conflict’ theory.

Although those who are poor are more exclusionist, the overall economic situation in the

area or its developments in the country seem to matter little, the same as the immigration

rates. It shows that exclusionist attitudes are more a matter of ideology.

MYPLACE August 2014


Exclusionism depends on the welfare state model: it is less widespread in the conservative

model (Germany) than in the liberal model (UK). From this point-of-view, the dismantling of

the European Social Model, observed during the recent economic crisis, risks facilitating the

spread of exclusionist attitudes among European youth.

Model E: Welfare chauvinism (Q41_2)

The initial analysis of variance components reveals that 79 percent of variance in welfare

chauvinism can be explained by differences between individuals, whereas 21 percent of the

variance lies at the level of localities (intra-class correlation (ICC) = 0.21) (Table 12.6.9). The

substantial variation at the level of localities means that it is necessary to employ a

multilevel analysis to try to find contextual variables responsible for it.


decrease of residual, individual level variance. Although the reduction might seem small, the

chi-square test of deviances (df = 25, loglikelihood = 2223) confirms that it is statistically

significant. Interestingly, adding individual level variables led to an even larger – 41 percent





containing only the individual level predictors. This means that our selected contextual

variables are capable of explaining a huge portion of the unexplained variation in welfare

chauvinism between localities. Only 5% of unexplained variance remains at the level of

localities, although one must note that it is still significant (at 0.01 level).

Table 12.6.9. Variance components of regression models explaining welfare chauvinism Empty

model

(random

intercept

only)

With

individual

level

explanatory

var’s

With

country

level

explanatory

var’s

σ (individual level) 1,057 ,973 ,973

Exclusionism σ (country level) ,279 ,164 ,051

-loglikelihood 41859,46 39636,56 39675,35

As with other variables measuring negative attitudes towards minorities, we find that ethnic

nationalism is among the strongest predictors of welfare chauvinism among youth (Table

12.6.10). A one-unit increase in ethnic nationalism on a 5-point scale is associated with a

0.235 unit increase in welfare chauvinism, measured on the same scale. One can conclude

that ethnic nationalism is also linked to welfare chauvinism.

MYPLACE August 2014


As in case of other variables we have looked at, welfare chauvinism is much more common

among men than women, as well as among young people with less than a tertiary-level

education, whose family is struggling financially, and whose parents do not belong to the

highest social class. Interestingly though those who are economically active (employed or

unemployed) are more exclusionist than students or those who are economically inactive. It

could be that they are trying to protect ‘what they have earned’, and not share it with

others who have not contributed to the countries’ welfare system themselves. Age is not

related to welfare chauvinism i.e., adolescents are less likely to hold such views as those

who are older.

Table 12.6.10. Determinants of exclusionism; n2=30, n1=14419 Model E

b SE


Gender (a man) ,059 *** ,017

Age ,000 ,004


Upper secondary education ,102 ** ,032

Less than upper secondary education ,052 * ,025

Employed Ref. Ref.

Unemployed ,004 ,038

In education -,070 * ,032

Other -,090 *** ,023

Class 3 Ref. Ref.

Class 2 ,104 *** ,024

Class 1 ,125 *** ,031

Class 0 ,054 ~ ,031

Financial problems in the household ,073 *** ,021

Dominant ethnicity in the country of residence ,044 ,032

Citizen of the county of residence ,204 *** ,046

Born in the county of residence ,059 ,043

Ethnic nationalism ,235 *** ,008

Political knowledge -,025 * ,010

Interest in politics -,046 *** ,011

Media exposure ,010 ** ,004

Social trust -,009 * ,004

Political trust -,022 *** ,004


Two friends of other ethnicity/race ,160 *** ,023

One friend of other ethnicity/race ,066 * ,028

No friends of other ethnicity/race ,065 * ,029


GDP growth during the crisis (2008-2011) ,006 ,006

% of families in the area experiencing financial difficulties ,015 *** ,004

% of young people born abroad -,002 ,007

% of young people of non-majority ethnicity -,006 ~ ,003

Net immigration rates ,082 * ,029

Size of the area (thousands) ,014 * ,005

MYPLACE August 2014


Liberal Ref. Ref.

Post-socialist -,635 * ,256

Nordic -,622 * ,251

Conservative -,645 * ,244

Mediterranean -,533 * ,240

Constant 1,487 *** ,294


Welfare chauvinism as regards to immigrants is more widespread among those who are

citizens of the country (citizenship is associated with 0.204 increase in welfare chauvinism),

although ethnic group and country of birth seem to have no effect. This indicates that

welfare chauvinism is based on competition for resources rather than cultural affinity.

In line with the ‘populist thesis’, the analysis confirms our expectations: young people who

are interested in politics, know more about it, and watch less media are less likely to hold

welfare chauvinist attitudes. Such attitudes are also linked to social and political distrust,

with those who have little trust in politicians and people in general being more welfare

chauvinist.

Welfare chauvinism is less widespread among those who have more contact with minorities,

especially if they have three or more friends of another race or ethnicity. The ‘contact

hypothesis’ is also supported by the fact that welfare chauvinism is lower in localities where

there are more people around who are of non-majority ethnicity (Sig.<0.1).

As regards to welfare chauvinism, we find support for the ‘realistic group conflict’ theory.

Such views are more common among the poorest youth and in localities where there are

many poor families, as well where there are high immigration rates, especially in larger

cities. Resource stress makes people more welfare exclusionist with regards to immigrants.

Welfare chauvinism also depends on the welfare state model: it is most widespread in the

liberal model (UK), which could be linked both to an ideological perspective, as well as to

weaker welfare provision and more insecurity in such welfare state model.

12.7 Summary

The MYPLACE survey shows that young people’s attitudes towards minorities differ from

country to country, and from locality to locality. Young people in Western European

locations, especially eastern Germany, are least likely to have negative views towards

different minority groups and to prefer limiting access by immigrants to a range of

resources. Young people in post-socialist locations, along with Greek locations, have more

negative attitudes and are more likely to be exclusionist towards immigrants. Nevertheless,

MYPLACE August 2014


in some countries there are large differences between localities, for example, in Latvia and

Finland.

The most negative attitudes are towards the Roma, particularly in in Hungary, Slovakia,

Greece and Russia where there is a considerable Roma minority, as well as Lieksa and

Nurmes (Finland). Attitudes towards Jews are more positive than negative. The most

negative young people can be found in Ozd (Hungary), Trnava (Slovakia), Aegyroupouli

(Greece), and Vyborg (Russia). Attitudes towards Muslims are similarly positive; however

there are localities where negative views towards them are quite widespread: Vyborg

(Russia), Trnava (Slovakia) and Sopron (Hungary).

Young people in Greek locations are much more likely than others to have xenophobic

attitudes; this should be seen in the context of Greece having taken a large proportion

those fleeing unrest from the ‘Arab Spring’. Limiting immigration would also be supported

by young people in Vyborg (Russia), Bareiro (Portugal), Trnava (Slovakia), and Nuneaton

(UK), whereas German locations are clustered at the bottom of the xenophobia scale.

Hungary, Russia and Georgia are at the top of the welfare chauvinism and exclusion scale;

here young people favour limiting access by immigrants to various resources. Interestingly,

Nuneaton youth in the UK tend to be much more xenophobic, welfare chauvinist and

exclusionist than youth in Coventry.

One of the strongest predictors of negative attitudes towards minorities, xenophobia,

welfare chauvinism and exclusionism is ethnic nationalism. Like other studies, we also find

that all these negative attitudes are more widespread among men, and among those who

are less educated, whose families are struggling financially and whose parents do not belong

to the highest social class. This can be linked to a higher level of insecurity and perceived

competition. Moreover, perhaps linked to that, we find that young people who are

employed are more likely to have such attitudes rather than those who are not employed.

Age, on the other hand, does not help to explain negative attitudes towards minorities;

contrary to psychological theories, we find that young people are not more likely to have

such negative views as those who are older.





towards minorities. This relationship is intriguing, and there could be several explanations

for it, such as cultural differences or perceived inter-group conflict or competition.



MYPLACE August 2014


to hold negative attitudes towards minorities. A frequent use of different media channels is,

however, strongly linked to negative attitudes towards minorities, xenophobia, welfare

chauvinism and exclusionism, which shows that the media are at least partly responsible for

the recent spread of populism and far right attitudes in Europe.

The data also confirms the ‘contact hypothesis’; personal contact with minorities decreases

negative attitudes towards minorities and immigrants. Similarly, overall distrust in people

and the government facilitates all kinds of negative attitudes towards minorities and

immigrants.


are poor also more xenophobic and likvely to have more negative attitudes towards

minorities, but the overall situation in the area matters too; in poorer areas such views are

more widespread. At the same time, the effect of GDP growth during the crisis is

insignificant. It shows that resource stress is mainly experienced based on the economic

conditions young people see around them, not general macro-economic developments.

Xenophobia and welfare chauvinism is more widespread where immigration rates are

higher, lending further support for resource stress and competition thesis. Finally, liberal

regimes are more likely to produce such attitudes. Thus the dismantling of the European



MYPLACE August 2014


References

Allport, G. W. (1954) The Nature of Prejudice. Cambridge, Mass.: Addison-Wesley.

Binder, N. E., Polinard, J. L., Wrinkle, R. D. (1997) Mexican American and Anglo attitudes

toward immigration reform: a view from the border, Social Science Quarterly 78 (2): 324–37.

Fetzer, J. S. (2000) Public Attitudes toward Immigration in the United States, France, and

Germany. New York: Cambridge University Press.

Hewstone, M., Rubin, M., Willis, H. (2002). Intergroup bias, Annual Review of Psychology 53:

575–604.

Hox, J.(2010) Multilevel Analysis, Techniques and Applications. Second edition. New York:

Routledge.

Mares, M. (2013) The extreme right's relationship with Islam and Islamism in East-Central

Europe: From allies to enemies, East European Politics and Societies, Online first: 1-20, DOI:

10.1177/0888325413502073

Yilmaz, F. (2012) Right-wing hegemony and immigration: How the populist far-right achieved

hegemony through the immigration debate in Europe, Current Sociology 60(3): 368– 381

Kääriäinen, J., Lehtonen, H. (2006). The variety of social capital in welfare state regimes – a

comparative study of 21 countries, European Societies 8 (1): 27–57.

Lewin-Epstein, N., Levanon, A. (2005) National identity and xenophobia in an ethnically

divided society, International Journal on Multicultural Societies 7(2):90-118.

Mau, S., Mewes, J., Schöneck, M. (2012) What determines subjective socio-economic

insecurity? Context and class in comparative perspective, Socio-Economic Review 10, 655–

682

Mudde, C. (2005) Racist Extremism in Central and Eastern Europe, East European Politics

and Societies 19 (2), 161–184.

Pettigrew, T. F. (1971) Racially Separate or Together? New York: McGraw-Hill.

Sherif, M., Sherif, C.W. (1953). Groups in Harmony and Tension. New York: Harper and

Brothers.

MYPLACE August 2014


Chapter 13: Democracy

Introduction

Respondents’ attitudes to democracy was measured using a range of questions covering

specific questions on satisfaction with democracy, types of political regime, through to the

legitimacy of violence, respect for human rights, experience of discrimination and attitudes

to social and economic issues.

Summary of National Reports

Country specific analyses were conducted as part of deliverable 4.5. This section briefly

summaries the key findings on understanding of democracy from these reports. To

generalise, most young people support democracy as a political system, but are less

satisfied with the way democracy works.

When asked questions about attitudes to various forms of political systems, most

respondents stated that, they were pro-democratic, supporting multi-party systems, with an

opposition that can freely express its views. A number of locations (Croatia, Georgia and to a

lesser extent the UK) had stronger autocratic views, including supporting having a strong

leader who is not constrained by parliament.

Satisfaction with democracy varied from country to country, ranging from a general

satisfaction with democracy (Denmark and western Germany) to dissatisfaction with

democracy (Greece, Spain and Portugal). Country specific analysis found that females were

more satisfied with democracy than males (Estonia, Finland, Latvia and Portugal),

satisfaction with democracy increased with social class (Denmark, Estonia, Finland and

Slovakia) and young people with higher education were more satisfied (Demark and western

Germany). Non-nationals were more satisfied with democracy than nationals (Spain and the

UK).

There was general disapproval for attitudes towards violence, with most young people

stating that it is rarely justified. Females are more against violence than males. A slightly

higher proportion of young people in a few countries (Croatia, Slovakia and the UK) stated

violence would be justified to protect human rights.

In general, there were low levels of self-perceived discrimination measured across

locations. There were higher levels of discrimination if the respondent was female (Finland,

MYPLACE August 2014


eastern and western Germany and Russia), were of an ethnic minority (Denmark, Estonia,

Finland, eastern Germany, Latvia and the UK), belonged to a particular religious minority

(Denmark, Latvia and UK). In certain countries, it was also perceived that specific

subcultures (Croatia and Russia) had higher levels of discrimination.

There were mixed views on the respect for Human rights. There is a general satisfaction

with human rights in Denmark, Finland and Germany (western and eastern), a polarised

view in Russia and a lack of respect in Greece and Georgia. There were also different

opinions between countries on the death penalty, with certain countries strongly

supporting the death penalty (Georgia). There are also polarised views (Estonia) and

generally more support from males than females.

Literature

Democracy is one of a number of types of regimes or systems of governance (Schmitter and

Karl, 1991). There is no consensus on what democracy is, and it is an essentially contested

concept (Coppedge, 2005). Democracy, understood in a very general way, means ‘rule by

the people’ (Coppedge et al., 2011), or ‘government by and for the people’ (Lijphart, 2012).

There are a number of forms of modern democracies, which, are sensitive to the historical

context from which they developed (Lijphart, 1999; Vatter and Bernauer, 2009).

Dahl (1971) states that there are essentially two attributes to a democracy, consisting of

contestation and participation. Supporting these are rights and laws, therefore, a

democracy system of government includes a number of key elements;

a) A system for choosing and replacing the government, through free and fair elections.

b) The active participation of the people, as citizens, in politics and civic life.

c) Protection of the human rights of all citizens.

d) A rule of law, in which the laws and procedures, apply equally to all citizens.

The two main types of democracy are ‘consensual’ and ‘majoritarian’; in majoritarian

systems, the majority rule, and exclude minority opinions in the decision making process,

whereas in a consensual system a broad range opinions or as many people as possible are

considered in decision making (Lijphart, 1999). Free and fair elections are held at regular

intervals on the basis of universal, equal and secret suffrage in all democratic systems

(Goodwin-Gill, 2006). Elections take place on matters that affect the lives of citizens (van der

Eijk and Franklin, 2009) and are contested on a range of economic, social and

communitarian issues (Coppedge, 2005). Democracies allow citizens to choose their leaders.

When a winning political party (or coalition of parties) takes office and are mandated to run

the country and when political decisions are made they, are supposed to ultimately to

reflect the wishes of a country’s citizens (van der Eijk and Franklin, 2009). Although there is

MYPLACE August 2014


a degree of uncertainty as to who will be elected and which policies will be pursued

(Schmitter and Karl, 1991). Governments may pursue either a socialist (social democratic) or

liberal (free-market) approach. Losing parties form the opposition, and on behalf of the

people, provide a challenge as part of an open and transparent political system. Within a

democracy, elections should provide an adequate representation, in terms of the spectrum

of views and policies (der Eijk and Franklin, 2009), from the left to the right, and should

provide openness and equity, with new entrants free to form new parties. Governments and

political parties (elites), are held to accountable by the people (citizens), between elections,

and through the electoral process they can either be re-elected or removed from office. This

process legitimises governments within democracies (Bogdanor, 2007). There have been a

number of waves of democracy (Schmitter and Karl, 1991), including the fall of the Berlin

Wall, in 1989, with Eastern European countries adopting democracies over communist

regimes.

Citizens are a key component within a democracy. For a democracy to work there must be

free, active participation of the people. This can take several forms, from the civic duty of

voting in elections, to voluntary participation in the ‘civil society’ (Schmitter and Karl, 1991)

which are non-governmental organisations and institutions, which involves freedom of

speech and expression, including protesting. A free and transparent media plays an

important role in communicating messages between the government, political parties

including the opposition and citizens.

Trust in government is a mainstay of democracy (Christensen and Laegreid, 2003) and

central to political trust is a positive appraisal of the performance of governments, parties

and leaders (Dermody et al., 2010). Government is assumed to possess the tools and

abilities to ‘solve’ social problems (Anderson and Gillory, 1997:68) and over recent years,

especially since the start of the economic recession, government effectiveness and output is

debated. Negative press coverage of political integrity and performance has led to a

negative orientation towards government and politicians, leading to cynicism and distrust

(Dermody et al., 2010:423). Over the last five years, trust in national governments (of the

EU) has fallen from 34% in autumn 2008, to 23% in autumn 2013; and trust in the European

Union has fallen from 47% to 31% over the same time period (European Commission, 2013).

In recent surveys, there has also been declining support for the broader European project

(Pew Research Center, 2013). There has been a fall in voter turnout for elections at national

and European levels; European election turnout has fallen from 62% participation in 1979,

to a disappointing 43% in 2009 (Nacarino et al., 2012). This together with a decline in

political efficacy or ’the feeling that individual political action does have, or can have an

impact on the political process‘ (Campbell et al., 1954:187) has led to what is described as a

‘democratic deficit’ (Norris, 2011, Kohler-Kohl, 2010, Bogdanor, 2007). The ‘democratic

deficit’ is when the perceived democratic performance diverges from public expectations

(Norris, 2011) with consequences leading to a crisis of legitimacy.

MYPLACE August 2014


Easton (1957) presents an approach to the analysis of political system, which is adopted by

Norris (2011), to develop a conceptual framework in an attempt to examine ‘democratic

deficit’. These including indicators of system support Norris (2011:24) which range from

diffuse (nation-state) to specific (incumbents) indicators;

1. Attitudes towards belonging to the nation-state e.g. national pride, citizenship

and identity

2. Agreement with core principles and normative values upon the regime is based

including support for democratic ideals and rejection of autocratic principles

3. Evaluations of the overall performance of the regime

4. Confidence in regime institutions

5. Approval of incumbent officeholders e.g. government leader, politicians

There are a number of methods for measuring democracies which are explored by Munck

and Verkuilen (2002), Coppedge (2005) and Coppedge et al. (2011), including Freedom

House Index (Freedom House, 2014), Economist Intelligence Unit Democracy Index

(Economist Intelligence Unit, 2013) and the Demos EU Democracy Index (Birdwell et al.,

2013). Each uses a range of variables to categorise a country’s political regimes into a series

of types. Likewise, there are similar indexes for Human Rights (Cingranelli and Richards,

2010) and Press Freedom (RSF, 2014), which complement broader democracy indices.

13.1 Satisfaction with democracy (Q42)

General satisfaction with the way that democracy works in each country is measured using

an eleven-point scale where high values indicate satisfaction with how well it is working.

MYPLACE August 2014


Q42: Satisfaction with democracy



provides a good eleven-point scale, where high values indicate

satisfaction with how well it is working

Graph 13.1: Mean satisfaction with democracy by location

Narrative

The overall mean for satisfaction with democracy is 5.01 (n=16,385,

sd=2.44). This varies by location, with the highest mean score at

Odense Center (Demark) with 7.43, and the lowest mean score at

New Philadelphia (Greece) with 3.34.


Finland, eastern and western Germany all in the top third, with higher

levels of satisfaction with democracy. Mediterranean countries

including Greece, Spain and Portugal clustered in the bottom third

with lower levels of satisfaction with democracy.

Most Countries had a similar mean score between the two locations;

notable exceptions were; Russia (Vyborg and Kupchino), 1.36

difference between two mean scores and Hungary (Downtown area of

Sopron and Downtown area of Ozd), 1.28 difference.

MYPLACE August 2014


13.2 Attitudes towards various forms of political systems

(Q43)

Respondents were asked about their views of different types of political systems in order to

assess how satisfied they are with the current system and how they might feel about

alternatives. Questions asked ‘very good’, ‘fairly good’, ‘neither good nor bad’, ‘fairly bad’

and ‘very bad’.

Q43_1: Having a strong leader who is not constrained by parliament

Q43_2: Having a democratic, multi-party system

Q43_3: Having the army rule

Q43_4: Having an opposition that can freely express its views

Further analysis of the data suggests that items under these two factors do have good

internal consistency to constitute two separate scales. Norris (2011) suggests that a

combining of Q43_2: ‘Having a democratic, multi-party system’ and Q43_4: ‘Having an

opposition that can freely express its views, supports the adherence to democratic values

and principles’. An ordinal-by-ordinal Gamma correlation of 0.525 suggests that there is a

strong relationship between the two variables and theoretically they can be combined.

Rejection of autocratic principles supports the adherence to democratic values and

principles. By combining Q43_1: ‘Having a strong leader who is not constrained by

parliament’ and Q43_3: ‘Having the army rule’, an index of autocratic principles can be

created. An ordinal-by-ordinal Gamma correlation of 0.521 suggests there is a strong

relationship between the two variables and theoretically, they can be combined.

MYPLACE August 2014


Q43: Democratic Views (Q43_2 & Q43_4)


An ordinal-by-ordinal Gamma correlation of 0.525 suggests that there

is a strong relationship between the two variables (Q43_2: ‘Having a

democratic, multi-party system’ and Q43_4: ‘Having an opposition

that can freely express its views’ and theoretically they can be

combined. The scale, which has a minimum of zero (negative views

towards democratic political systems) to 8 (positive views towards

democratic political systems)

Graph 13.2a: Mean views towards Democratic Political Systems by

location

Narrative

The overall mean for views towards democratic political systems is

5.83 (n=14,693, sd=1.53) out of a possible score of 8. This varies by

location, with the highest mean score at Odense Center (Denmark)

with 6.92, and the lowest mean score in Vyborg (Russia) with 4.76.

There are clear clusters of locations; with locations in Denmark and

eastern Germany all in the top third, with more positive views

towards democratic political systems. Locations in Russia, Latvia and

Slovakia are clustered in the bottom third.

Most Countries had similar mean scores between the two locations;

notable exceptions were; Finland (Kuopio and Lieksa and Nurmes),

0.674 difference between two mean scores, and Greece (New

Philadelphia and Argyroupouli), 0.460 difference.

MYPLACE August 2014


Q43: Autocratic Principles (Q43_1 & Q43_3)


An ordinal-by-ordinal Gamma correlation of 0.521 suggests that there

is a strong relationship between the two variables (Q43_1: ‘Having a

strong leader who is not constrained by parliament’ and Q43_3:

‘Having the army rule’) and theoretically they can be combined. The

scale, which has a minimum of zero (negative autocratic principles) to

8 (positive autocratic principles).

Graph 13.2b: Mean views towards Autocratic principles by location

Narrative

The overall mean for views towards democratic political systems is

4.70 (n=15,377, sd=1.49) out of a possible score of 8. This varies by

location, with the highest mean score at Kutaisi (Georgian) with 5.75,

and the lowest mean score in Bremen (western Germany) with 3.76.


Croatia, UK and Portugal all in the top third, with more positive views

towards autocratic principles. Locations in Germany (western and

eastern), Finland and Spain are clustered in the bottom third with

negative views towards autocratic principles.


notable exceptions were; Hungary (Ozd and Sopron), 0.584 difference

between two mean scores, Greece (New Philadelphia and

Argyroupouli), 0.490 difference, and Estonia (Narva area and Tartu),

0.455 difference.

MYPLACE August 2014


13.3 Attitude towards violence (Q44)

Respondents were asked in which of the following cases do you believe violence can be or

cannot be justified?’ with the following options; ‘always justified’, ‘mostly justified’, ‘neither

justified or unjustified, mostly unjustified and never justified

To protect jobs from being cut

To protect human rights

To prevent global warming

To end poverty

To protect your ethnic / racial group

To sustain a stable government

To overthrow a government

To respect and protect animal rights

MYPLACE August 2014


Q44: Justification for violence


A factor analysis of these items showed that all eight items are loaded

under a single factor with excellent reliability (α = 0.928). We

therefore constructed an index composed of responses from each

item and analysed this scale by location. The scale, which has a

minimum of zero (‘violence never justified’) to 32 (‘violence always

justified’). Individual country Cronbach alpha varied between 0.864 in

eastern Germany and 0.967 in Hungary.

Graph 13.3: Mean Justification for violence by location

Narrative

The overall mean for justification for violence is 9.72 (n=15,664,

sd=7.99) out of a possible 32. This varies by location, with the highest

mean score at Rimavska Sobota (Slovakia) with 14.36, and the lowest

mean score at Forstate & Jaunbuve (Latvia) with 4.62.

There are clear clusters of locations; with locations in Russia, Croatia,

and UK all in the top third, stating that violence is ‘more justified’.

Locations in Latvia, Denmark, Finland and Estonia are clustered in the

bottom third with stating that violence is ‘less justified’.


notable exceptions were; Slovakia (Rimavska Sobota and Trnava), 4.14

difference between two mean scores, Latvia (Agenskalns apkaime in

Riga and Forstate & Jaunbuve), 3.28 difference, and Finland (Lieksa

and Nurmes and Kuopio), 2.64 difference.

MYPLACE August 2014


13.4 Attitude towards social justice and the economy (Q45)

Respondents’ attitudes towards social justice and the economy were measured by a four-

item question, which measures differing views on the economy from a variety of

perspectives. The items are balanced with two which are ideologically closer to socialist

principles (items 2 and 3), and two which are closer to free-market principles (items 1 and

4). This structure allows us to explore the individual ideological positions held and the

extent to which they can be mapped onto an ideological spectrum.

Q45_1 The unemployed should have to take any job available or lose their unemployment benefits

Q45_2 Competition is harmful, it brings out the worst in people

Q45_3 Incomes should be made more equal

Q45_4 Private ownership of business and industry should be increased

The factor analysis of these four items showed there to be two distinct factors, which

mapped directly onto the item pairing described above. However, the reliability of these

factors was low with Cronbach α values less than 0.4. For this reason, we did not produce

indices and analyse each of the items separately.

MYPLACE August 2014


Q45_1: The unemployed should have to take any job available or lose their unemployment benefits



responses.

Graph 13.4a: The unemployed should have to take any job available

or lose their unemployment benefits by location

Narrative



disagree’, 28% ‘disagree’ and 8% ‘strongly disagree’, that the

unemployed should take any job available or lose their

unemployment benefits by location.

Trnava (Slovakia) had the largest proportion of respondents (58%)

strongly agreeing or agreeing. Telavi (Georgia) had the smallest

proportion (23%).

Tartu (Estonia) had the largest proportion of respondents (37%) who

neither agree or disagree.

There are less clustering of locations by country, although both

locations in Denmark are in the top third. When ordered by ‘disagree’

and ‘strongly disagree’, both locations in Georgia, Greece and eastern

Germany are in the top third when stating that the unemployed

should take any job available or lose their unemployment benefits

MYPLACE August 2014


Q45_2: Competition is harmful, it brings out the worst in people



responses.

Graph 13.4b: Competition is harmful, it brings out the worst in

people by location

Narrative



disagree’, 38% ‘disagree’ and 10% ‘strongly disagree’, that

‘Competition is harmful, it brings out the worst in people’.

Argyroupouli (Greece) had the largest proportion of respondents

(45%) ‘strongly agreeing or agreeing’. Odense Center (Denmark) had

the smallest proportion (7%).

There are clear clusters of locations, with locations in Greece, Spain,

Hungary, Russia and Slovakia all in the top third, ‘agreeing’ that

competition is harmful. When ordered by ‘disagree’ and ‘strongly

disagree’, locations in Denmark and Germany (western and eastern),

are clustered in the top third ‘disagreeing’ that competition is

harmful.

MYPLACE August 2014


Q45_3: Incomes should be made more equal



responses.

Graph 13.4c: Incomes should be made more equal by location

Narrative



disagree’, 14% ‘disagree’ and 3% ‘strongly disagree’, that ‘Incomes

should be made more equal’.

New Philadelphia (Greece) had the largest proportion of respondents

(86%) ‘strongly agreeing or agreeing’. Odense Center (Denmark) had

the smallest proportion (23%).

There are clear clusters of locations, with locations in Greece, and

Germany (eastern and western) all in the top third, ‘agreeing’ that

Incomes should be made more equal. When ordered by ‘disagree’ and

‘strongly disagree’, locations in Denmark, Georgia and Estonia are

clustered in the top third disagreeing that Incomes should be made

more equal.

MYPLACE August 2014


Q45_4: Private ownership of business and industry should be increased



responses.

Graph 13.4d: Private ownership of business and industry should be

increased by location

Narrative



disagree’, 20% ‘disagree’ and 5% ‘strongly disagree’, that ‘Private

ownership of business and industry should be increased’.

Telavi (Georgia) had the largest proportion of respondents (95%)

‘strongly agreeing or agreeing’. Jena (eastern Germany) had the

smallest proportion (12%).

There are clear clusters of locations, with locations in Georgia and

Latvia in the top third, agreeing that Private ownership of business

and industry should be increased. When ordered by ‘disagree’ and

‘strongly disagree’, locations in eastern Germany and Greece are

clustered in the top third ‘disagreeing’ that Private ownership of

business and industry should be increased.

MYPLACE August 2014


13.5 Freedom of speech (Q46)

Q46_1 Holocaust denial

Q46_2 Islamic extremism

Q46_3 Anti Roma / Traveller sentiment

Respondents’ views on freedom of speech was measured using three 0-10 scales asking

about people being allowed to speak in public in support of Holocaust denial, Islamic

extremism and anti-Roma sentiment. These items represent contrasting issues with

different sensitivities in different parts of Europe.

There are issues with missing values in some countries, which is likely to be due to

contemporary experience of specific groups.

In Georgia 38% of cases are missing for this derived scale. There are 27% of missing cases in

Finland and Russia, 23% in Greece and 22% in Hungary.

The table below illustrates the proportion of missing cases by the three questions and the

derived variable.

Table 13.5 Proportion of Missing Values by question

Country

Derived

variable

Holocaust

denial

Islamic

extremism

Anti Roma

/ Traveller

sentiment

Croatia 18 16 16 14

Denmark 8 6 2 3

Estonia 20 17 15 14

Finland 27 24 21 19

Georgia 38 35 29 26

Western Germany 13 5 3 10

Eastern Germany 11 2 2 9

Greece 23 20 18 13

Hungary 22 20 21 19

Latvia 9 8 7 6

Portugal 19 17 15 10

Russia 27 25 19 18

Slovakia 19 17 16 12

Spain 11 10 7 6

UK 11 7 4 6

Total 19% 15% 13% 12%

MYPLACE August 2014


Q46: Freedom of speech


A factor analysis of these items showed that all three items are loaded

under a single factor with excellent reliability (α = 0.866). We

therefore constructed an index composed of responses from each

item and analysed this scale by location. The scale, which has a

minimum of zero (‘should never be allowed’) to 30 (‘should always be

allowed’). Individual country Cronbach alpha varied between 0.736 in

Georgia and 0.898 in Portugal.

Graph 13.5: Mean Freedom of speach by location

Narrative

The overall mean for freedom is 11.07 (n=13,788, sd=8.14). This varies

by location, with the highest mean score at Barreiro (Portugal) with

17.61, and the lowest mean score at Bremen (western Germany) with

3.09.

There are clear clusters of locations; with locations in Portugal,

Denmark, Greece and Estonia all in the top third, stating that freedom

of speech should always be allowed. Germany (western and eastern),

Latvia and Hungary are clustered in the bottom third.

Most Countries had similar a mean score between the two locations;

notable exceptions were; Finland (Lieksa & Nurmes and Kuopio), 3.65

difference between two mean scores and Latvia (Agenskalns apkaime

in Riga and Forstate & Jaunbuve), 2.36 difference.

MYPLACE August 2014


13.6 Respect for human rights (Q47)

Respondents were asked for their views on how much respect is there for individual human

rights in their country; with the following options; ‘a great deal of respect’, ‘a fair degree of

respect’, ‘not much respect’ and ‘no respect at all’.

MYPLACE August 2014


Q47: Respect for human rights


The graph is ordered by combining the ‘a great deal of respect’ and

the ‘a fair degree of respect’ responses.

Graph 13.6: Respect for human rights by location

Narrative


respondents stating ‘there is a great deal of respect’ for human right,

49% said a ‘fair degree of respect’, 31% said ‘not much respect ‘and

7% said ‘no respect at all’.

Kuopio (Finland) had the largest proportion (97%) stating that ‘there is

respect for human rights’. Telavi (Georgia) had the smallest

proportion of respondents (29%) stating ‘there is respect for human

rights’.

There are clear clusters of locations; with locations in Finland,

Denmark and Germany (eastern and western) are all in the top third,

stating that there is respect for human rights. Both locations in

Georgia, Greece and Slovakia are in the bottom third when stating

there is respect for human rights.

MYPLACE August 2014


13.7 Self-perceived discrimination (Q48)

Self-perceived discrimination or Sense of Insecurity in Society (Q48) contains the following

variables;

Q48_1: you support a particular political movement

Q48_2: you are a member of an ethnic or religious minority

Q48_3: of your sexual orientation

Q48_4: you belong to a subculture (punk / skinhead / EMO / Goth etc.)

Q48_5: of your gender

These five items were used for measuring self-perceived discrimination/Sense of Insecurity

in Society index. Young people responded to each item (representing a particular aspect of

discrimination) on a three-point scale: ‘never’, ‘occasionally’, and ‘regularly’. In constructing

an index, respondents who either reported ‘occasionally’ or ‘regularly’ for each item are

counted, which resulted in an index ranging from zero to five. Higher scores in the index

indicates greater levels of discrimination.

MYPLACE August 2014


Q48: Self-perceived discrimination/Sense of Insecurity in Society


An index ranging from zero to five, where a higher score in the index

indicates a greater level of discrimination/sense of insecurity in

society.

Graph 13.7: Mean of Sense of Insecurity in Society Index by location

Narrative

Overall levels of discrimination are low. The overall mean for sense of

insecurity in society is 0.367 (n= 16227, sd=0.802). This varies by

location, with the highest mean score at Podsljeme (Croatia) with

0.749, and the lowest mean score at Barreiro (Portugal) with 0.125.

There are clear clusters of locations, with locations in Croatia, UK,

Russia, eastern Germany, and Slovakia all in the top third, with higher

levels of sense of insecurity/ discrimination. Portugal, Hungary and

Georgia are clustered in the bottom third with lower levels of sense of

insecurity/discrimination.

MYPLACE August 2014


13.8 Opinion on death penalty (Q49)

Young People were asked is they ‘think that the death penalty can sometimes be justified or

do you believe that the death penalty can never be justified?’ with measured using an

eleven-point scale where high values indicate that the death penalty can never be justified.

MYPLACE August 2014


Q49: Opinion on death penalty



provides a good eleven-point scale, where high values indicate the

belief that the death penalty can never be justified.

Graph 13.8: Mean of Justiification of Death Penalty by location

Narrative

The overall mean for justification of the death penalty is 5.64

(n=16,569 sd=3.70). This varies by location, with the highest mean

score at Kutaisi (Georgia) with 8.68, and the lowest mean score at Ozd



Denmark and Greece all in the top third, stating that the death

penalty can never be justified. Locations in Hungary, Estonia, Latvia,

and Slovakia are clustered in the bottom third with more respondents

stating that the death penalty can be justified.

Most Countries had similar a mean score between the two locations;

notable exceptions were; western Germany (Bremen and

Bremerhaven), 1.65 difference between two mean scores and Russia

(Part of Vyborg and Kupchino), 1.41 difference.

MYPLACE August 2014


of 431


This section presents the results of multi-level modelling. Multi-level modelling (MLM) was

used to take into account the nested structure (respondents within locations within

countries) in our dataset.

Model: Satisfaction with democracy

Dependent variable

The dependent variable used is the ‘satisfaction with democracy’ scale from question 42. As

this variable is an almost normally distributed scale variable, a linear mixed model was used.


Our selection of independent variables, in the model, is guided by existing theories and/or

previous studies on ‘satisfaction’ with democracy. The literature review identified two

dominant theories on people's satisfaction with the way democracy works: (a) Cultural and

historical theories, and (b) system performance theories. Cultural and historical theories

argue that the differences in ‘satisfaction’ with democracy can be explained by the relative

presence or absence of a civic (or democratic political) culture (Almond and Verba, 1965;

Inglehart, 1988; Weil, 1989). They consider that a democratic political culture is the result of

distinct national heritages, traditions and historical legacies, at the level of countries, and

distinct socialisation experiences at the level of individual citizens (Inglehart, 1990). Guided

by these theories, we include two independent variables (one contextual and one individual

level) in our model on satisfaction with democracy: the Welfare State type (Kaariainen and

Lehtonen 2006) [contextual variable at level 2] : five categories (i) Post-socialist, (ii) Nordic,

(iii) Conservative, (iv) Mediterranean and (v) Liberal, and Interpersonal trust. In order to

measure whether a democratic political culture exists at the level of individual citizens and

how firmly it has been entrenched, we have used interpersonal trust as an individual level

indicator of the civic culture syndrome (Anderson, 2001).

On the other hand, system performance theorists argue that people's satisfaction with

democracy depend on what kind of outputs it produces (Easton, 1975). They have identified

apparent performance-related weaknesses, as important influences on citizen

dissatisfaction with democracies (Brittan, 1975; Huntington, 1974; Barnes and Kaase 1979;

Jennings and van Deth, 1989). Influenced by theory, Anderson (2001) focused on political

and economic performance in explaining satisfaction with democracy. In our analysis, we

also include these two variables. To measure political performance of a country [level 1

variable], young people's views, regarding the country's level of respect for individual

human rights, were gathered on a four-point scale ranging from 0 (‘no respect’) to 3 (‘a

MYPLACE August 2014


of 431

great deal of respect’ for individual human rights). To measure economic performance

related weakness of a country (level 2 variable), percentage of young people unemployed

in the country is used in this study.

Apart from these variables, literature on satisfaction with democracy identifies a number of

other variables as correlates. For example, Klingemann (1999) finds that satisfaction with

democracy goes along with higher levels of support for democracy. Therefore, young

people's views on 'having a democratic, multi-party system' as a form of government

(political system)’ is included in the analysis. This question had five response options: ‘very

bad’ (0), ‘fairly bad’ (1), ‘neither good nor bad’ (2), ‘fairly good’ (3), and ‘very good’ (4). A

higher score indicates greater support for democracy.

Previous studies also found trust in political institutions as a significant correlate of citizens'

satisfaction with democracy (Zmerli et al.,2007). For measuring trust in political institutions,

we asked young people’s level of trust in parliament in an eleven-point scale ranging from

zero (‘do not trust at all’) to 10 (‘complete trust’). Trust in political actors (Norris 1999) is

also linked with people's satisfaction with democracy (Schafer, 2013; Nevitte and Kanji,

2002). In order to measure young people's level of trust for politicians, we used the

statement 'Politicians are corrupt'. Respondents expressed their views on this statement in

a five-point scale ranging from 0 (‘strongly agree’) to 4 (‘strongly disagree’). A higher score

indicates greater level of trust for politicians.

In relation to interest in politics, literature has demonstrated that it is related to political

efficacy and political support (Almond and Verba 1965; Lambert et al., 1986; Weatherford,

1991). Interest in politics is also found to be associated with satisfaction with democracy

(Schafer, 2013; Anderson and Guillory, 1997). Young people's level of interest in politics is

measured in a four-point scale ranging from 0 (‘not at all interested’) to 3 (‘very interested’).

In addition, citizens who are knowledgeable about the political process are likely to have a

more optimistic view of democratic governance. Political knowledge is measured in the

study by counting the correct responses (correct/incorrect) to each political knowledge

question (‘who is…head of government, foreign minister, ruling party’). Putnam (1993,

1995, 2000) argues that weak social capital is linked with lower satisfaction with democracy.

Participation is now widely recognised as a measure of social capital (Nevitte and Kanji,

2002). For measuring participation, we have developed a Political activism Index by

counting those activities (from a list of 20, including volunteering in an election campaign,

signing a petition, wearing a badge with a political message, etc.) which the young people

were involved in at least once during the last 12 months.

Democratic theory argues that democracy enhances the quality of life for citizens in

fundamental ways, because the very freedoms that make democratic contestation and

participation possible, also enable citizens to pursue various private goals directly and

MYPLACE August 2014


of 431

indirectly (Dahl, 1989; Pateman, 1970). Furthermore, responsive democracy is likely to

implement social policies that enhance the well-being, or at least the security of large

numbers of people. As Esping-Anderson (1990) notes, a democratic state is by necessity

preoccupied with the production and distribution of social well-being. Frey and Stutzer

(2000) examine the link between happiness and direct democracy in Switzerland from 1992-

1994. They find that direct democracy has a statistically significant, positive effect on

happiness. Moreover, as the level of direct democracy increases, in the 26 Swiss cantons,

levels of happiness correspondingly increase. In our model, we include satisfaction with life

as a correlate of ‘satisfaction’ with democracy. In relation to minority groups who live in

multi-ethnic countries, Ruiz-Rufino (2013) identified that the regime from which individuals

originate, will condition their attitudes about democracy even if the regime alters. This may

also be applied if they migrate to a new country. Therefore we include identity measured as

0 (‘minority in country’) and 1 (‘majority in country’) in the model.

Socio-economic resource theory (Schafer, 2013) maintains that satisfaction with democracy

is positively linked with socio-economic resources. Influenced by this theory, we include

subjective assessment of household income, unemployment status, and parental class as

three correlates of satisfaction with democracy. Among other socio-demographic

characteristics, Schafer (2013) found gender and age as two significant correlates of

satisfaction with democracy. Sanders et al. (2014) identified that specific types of religion

were significant in their model, with Catholic and Protestants more satisfied with democracy

than their non-religious counterparts were. Therefore we include religion measured as 0

(‘not religious’) and 1 (‘religious’) in the model.

Results

The initial analysis of variance components in the ‘empty’ model in Table 13.8a reveals that

77.5% of the variance in young people’s ‘satisfaction’ with democracy can be explained by

differences between individuals. Furthermore, an intra-class correlation of 0.225 suggests

that 22.5 percent of the overall variance in the ‘satisfaction’ with democracy lies at the level

of localities. The substantial variation at local level justifies the reason for using multilevel

analysis in our data.

The inclusion of the described individual variables (in Model 1 in Table 13.8.1) resulted in a

28.3% decrease of residual, individual level variance. Chi-square test of deviances

(loglikelihood = 4336.484; df = 13) confirms that this is a significant difference. Adding

individual level variables, led to a 43.3% reduction in the intercept variance, indicating that

the amount of variance of ‘satisfaction’ with democracy between localities can be attributed

to individual level predictors.

In the next model (Model 2 in Table 13.8.a), we included contextual variables which led to a

73.5% reduction of intercept variance, compared to the model containing only individual

MYPLACE August 2014


of 431

level predictors. Chi-square test of deviances (loglikelihood = 32.32; df = 5) confirms that

this is a significant difference. The ICC value of .0573 in the final model indicates that almost

6% of the unexplained variance remains at the level of localities. In other words, there is a

scope for improving the model by including other relevant variables in the analysis.

Table 13.8.1. Variance components of regression models explaining satisfaction with

democracy Empty model

(random

intercept only) -

Model 0

With individual

level

explanatory

var’s – Model 1

With country

level

explanatory

var’s – Model 2

Satisfaction with

democracy


σ (location level) 1.360 0.771 0.204

-2 log likelihood 58203.482 53866.998 53834.675

df 3 19 24

Table 13.8.2: Determinants of Satisfaction with Democracy; n2=30, n1=13,252

Model 2

Individual-level predictors b

SE

Unemployed .079 (0.056)

Employed Ref. Ref.

Gender (male) -.162 *** (0.032)

Female Ref. Ref.

In Minority .115 * (0.048)

In Majority Ref. Ref.

Not Religious -.272 *** (0.041)

Religious Ref. Ref.

Social Class -.035 * (0.014)

Age -.019 ** (0.006)

Political Knowledge .096 *** (0.020)

Life Satisfaction .155 *** (0.009)

Trust in Parliament .260 *** (0.008)

Household Income .030 (0.022)

Interest In politics -.024 (0.021)

Trust in Politicians .245 *** (0.018)

Attitudes to democratic, multi-party system .237 *** (0.019)

Political Activism Index -.052 *** (0.006)

Human rights .696 *** (0.024)

General Trust in People .020 ** (0.007)


Nordic 1.304 *** (0.310)

Conservative 1.192 ** (0.365)

Mediterranean -.574 (0.322)

Liberal .297 (0.361)

Post-socialist Ref. Ref.

Youth Unemployment -.009 (0.012)

Constant 3.072 *** (0.380)

***p<0.001, **p<0.01, *p<0.05; two-tailed tests.

MYPLACE August 2014


of 431

Table 13.8.2 presents results on the relationships of all individual and contextual level

variables with young people’s ‘satisfaction’ with democracy. Among 16 individual level

variables, included in the model, unemployment status, household income situation, and

interest in politics, were not significantly associated with young people’s ‘satisfaction’ with

democracy. However, controlling the effects of both individual and contextual variables in

the model, young males, those with greater involvement with political activities, older

young people, those having parents from upper social class, and those ‘not religious’,

appeared to have statistically lower ‘satisfaction’ with democracy. In addition, young people

reporting greater political knowledge, satisfaction with life, trust in parliament, trust in

politicians, general trust in people, and positive evaluation of country’s human rights

situation have statistically significant higher level of ‘satisfaction’ with democracy.

Compared to their majority counterparts, young people from minority backgrounds are also

reported to have greater ‘satisfaction’ with democracy in the country.

Controlling the effects of those individual level variables, and welfare state type in the

model (Table 13.8.2); youth unemployment of the country did not appear to have a

statistically significant relation with young people’s ‘satisfaction’ with democracy. However,

welfare state type did have a significant association with ‘satisfaction’ with democracy.

Compared to those from post-socialist countries, young people from Nordic and

Conservative type countries appeared to have significantly greater ‘satisfaction’ with

democracy.

13.9 Summary

In this chapter, we reported the findings of cross-country analyses of the MYPLACE survey,

in which young people’s understandings of democracy are analysed.

A descriptive analysis of Satisfaction with democracy shows that the overall mean across all

30 locations is exactly in the middle of the scale (5.0), but satisfaction varies from country to

country and between locations. There were clear clusters of locations; with locations in

Denmark, Finland, eastern and western Germany all having higher levels of satisfaction with

democracy. Mediterranean countries including Greece, Spain and Portugal have lower levels

of satisfaction with democracy.

An ordinal by ordinal Gamma correlation of Attitudes towards various forms of political

systems (Q43) suggested that two separate scales were formed, one which focused on

Democratic Political System and another which focused on Autocratic principles. Results

showed that young people in locations in Denmark and eastern Germany had more positive

views towards democratic political systems. Respondents in locations in Russia, Latvia and

MYPLACE August 2014


of 431

Slovakia were less positive towards democratic political systems. Both locations in Georgia,

Croatia, the UK and Portugal have more positive opinions towards a non-democratic form of

political system (with autocratic principles ). Both locations in Germany (western and

eastern), Finland and Spain were more negative towards autocratic principles.

A factor analysis showed that all eight items for question 44: Justification for violence are

loaded under a single factor with excellent reliability (α = 0.928). The overall mean for

justification for violence was generally low (9.7 out of a possible 32). Both locations in

Russia, Croatia, and the UK were in the top third, stating that violence is ‘more justified’,

with locations in Latvia, Denmark, Finland and Estonia stating that violence is ‘less justified’.

Items in question 45: Attitude towards social justice and the economy were analysed

individually, because when combined the reliability of these factors was low, with Cronbach

alpha values less than 0.4. This suggests that there are contradictions between respondents’

ideological views in terms of socialist principles (items 2 and 3), and free-market principles

(items 1 and 4). Both locations in Georgia, Greece and eastern Germany stated that they

disagreed that the unemployed should have to take any job available or lose their

unemployment benefits. Both locations in Portugal, Denmark and Hungary are in the top

third and stated that they should. Most young people in the two Georgian locations stated

that Private ownership of business and industry should be increased. Both locations in

eastern Germany and Greece are clustered in the top third disagreeing that Private

ownership of business and industry should be increased. There are clusters of locations,

with locations in Greece, Spain, Hungary, Russia and Slovakia all in the top third, agreeing

that Competition is harmful, it brings out the worst in people. When ordered by ‘disagree’

and ‘strongly disagree’, locations in Denmark and Germany (western and eastern) are

clustered in the top third disagreeing that competition is harmful. There are clear clusters of

locations, with young people’s views on Incomes should be made more equal, with

locations in Greece, and Germany (eastern and western) all in the top third, agreeing that

incomes should be made more equal. When ordered by ‘disagree’ and ‘strongly disagree’,

locations in Denmark, Georgia and Estonia are clustered in the top third disagreeing that

Incomes should be made more equal.

A factor analysis of these items showed that all three items are loaded under a single factor

with excellent reliability (α = 0.866) for Freedom of speech. Both locations in Portugal,

Denmark, Greece and Estonia all in the top third, stating that freedom of speech should

always be allowed. Germany (western and eastern), Latvia and Hungary are clustered in the

bottom third.

There is a large variation of views regarding respect for human rights. Both locations in

Finland, Denmark and Germany (eastern and western) are all ranked in the top third, stating

that there is respect for human rights. Both locations in Georgia, Greece and Slovakia are in

MYPLACE August 2014


of 431

the bottom third when stating there is no respect for human rights. Both locations in Greece

had the highest proportion of young people stating there is ‘no respect at all’ (22% and

24%).

An index ranging from zero to five, where a higher score in the index indicates greater level

of discrimination/sense of insecurity in society, were constructed for the five items in the

question on Self-perceived discrimination (Q48). Overall levels of discrimination are low;

with a mean for sense of insecurity in society of 0.4. When ranked, locations in Croatia, the

UK, Russia, eastern Germany, and Slovakia are all in the top third, with higher levels of sense

of insecurity/ discrimination. Portugal, Hungary and Georgia are clustered in the bottom

third with lower levels of sense of insecurity/discrimination.

Opinions on the death penalty (Q49) vary from country to country, with locations in

Georgia, Denmark and Greece all in the top third, stating that the death penalty can ‘never

be justified’. However, Locations in Hungary, Estonia, Latvia, and Slovakia are clustered in

the bottom third with more respondents stating that the death penalty ‘can be justified’.

In order to explore multivariate determinants of Satisfaction with democracy, taking into

account the nested structure in our data set, locations within countries, we employed

multilevel modelling (MLM). The initial analysis of variance components in the ‘empty’

model revealed that 78% of the variance in young people’s ‘satisfaction’ with democracy

can be explained by differences between individuals and that 23% of the overall variance, in

the ‘satisfaction’ with democracy, lies at the level of localities. The substantial variation at

local level justifies the reason for using multilevel analysis in our data. The model confirmed

that young people reporting greater political knowledge, satisfaction with life, trust in

parliament, trust in politicians, general trust in people, and positive evaluation of the

country’s human rights situation have a statistically significant higher level of ‘satisfaction’

with democracy. Young males, those with greater involvement with political activities, older

young people, those having parents from upper social class, and those ‘not religious’

appeared to have a statistically lower level of satisfaction with democracy. At a contextual

level, the youth unemployment rate of a country did not appear to have a statistically

significant relationship to young people’s ‘satisfaction’ with democracy, but welfare state

type did have a significant association with ‘satisfaction’ with democracy. Compared to

those from post-socialist countries, young people from Nordic and Conservative type

countries appeared to have significantly greater ‘satisfaction’ with democracy.

MYPLACE August 2014


of 431

References

Almond, G. A., and Verba, S. (1965). The Civic Culture. Boston: Little, Brown.

Anderson, C. J. (2001). Political Satisfaction in Old and New Democracies. Binghamton: NY:

Binghamton University. Centre on Democratic Performance. Working Paper No. 102.

Anderson, C. J., and Guillory, C. A. (1997). Political Institutions and Satisfaction with

Democracy: A Cross-National Analysis of Consensus and Majoritarian Systems. The American

Political Science Review. 91 (1): pp. 66-81.

Barnes, S., and Kaase, M. eds. (1979). Political Action: Mass Participation in Five Western

Democracies. Beverly Hills, CA: Sage.

Birdwell, J., Feve, S., Tryhorn, C. and Vibla, N. (2013) Backsliders: Measuring Democracy in

the EU. Demos

Brittan, S. (1975). The Economic Contradictions of Democracy. British Journal of Political

Science. 5 (April):129-59.

Bogdanor, V. (2007) Legitimacy, Accountability and Democracy in the European Union. A

Federal Trust report

Christensen, T. and Laegreid, P. (2003) Trust in Government – the Significance of Attitudes

Towards Democracy, the Public Sector and Public Sector Reforms.

Cingranelli, D. L. and Richards, D. L. (2010) The Cingranelli and Richards (CIRI) Human Rights

Data Project. Human Rights Quarterly 32: pp. 395-418

Coppedge, M. (2005) Defining and Measuring Democracy. Political Concepts Committee on

Concepts and Methods Working Paper Series 2 pp. 1-57

Coppedge, M. and Gerring, J. with Altman, D., Bernhard, M., Fish, S., Hicken, A., Kroenig,

M., Lindberg, S. I., McMann, K., Paxton, P., Semetko, H. A., Skaaning, S-E., Staton, J. and

Teorell, J. (2011) Conceptualizing and Measuring Democracy: A New Approach Perspective

on Politics (9)2: pp. 247-267

Dahl, R. A. (1971) Polyarchy: Participation and Opposition. New Haven: Vale University

Press.

Dahl, R. A. (1989). Democracy and its Critics. New Haven/London: Yale University Press.

Dermody, J., Hanmer-Lloyd, S. and Scullion, R. (2010) Young people and voting behaviour:

alienated youth and (or) an interested and critical citizenry? European Journal of Marketing

44 (3/4): pp. 421-435

Easton, D. (1957) An Approach to the Analysis of Political Systems. World Politics 9(3): pp.

383-400

MYPLACE August 2014


of 431

Easton, D. (1975). A Re-Assessment of the Concept of Political Support. British Journal of

Political Science. 5 (October) pp. 435-37.

Economist Intelligence Unit (2013) Democracy index 2012: Democracy at a standstill

http://www.eiu.com/public/topical_report.aspx?campaignid=DemocracyIndex12

Esping-Anderson, G. (1990). The Three Worlds of Welfare Capitalism. Princeton: Princeton

University Press.

European Commission (2013) Standard Euroabarometer 80: Public Opinion in the European

Union

Freedom House (2014) Freedom in the World 2014

Frey, B. and Stutzer, A. (2000). “What are the Sources of Happiness?” A working paper

available on-line at http://www.economics.uni-linz.ac.a.

Goodwin-Gill, G. S. (2006) Free and Fair Elections. Geneva: Inter-Parliamentary Union

Huntington, S. (1974). Postindustrial Politics: How Benign Will It Be? Comparative Politics, 6

(January) pp. 163-91.

Inglehart, R. (1988). The Renaissance of Political Culture. American Political Science Review.

82 (December) pp. 1203-1232.

Inglehart, R. (1990). Culture Shift in Advanced Industrial Society. Princeton, NJ: Princeton

University Press.

Jennings, M. Kent, and van Deth, J. eds. (1989). Continuities in Political Action. New York: de

Gruyter.

Kääriäinen, J., Lehtonen, H. (2006). The variety of social capital in welfare state regimes – a

comparative study of 21 countries, European Societies 8 (1) pp. 27–57.

Klingemann, Hans-Dieter. (1999). Mapping Political Support in the 1990s: A Global Analysis.

In: Pippa Norris (ed.) Critical Citizens. Global Support for Democratic Government (pp. 31-

56). Oxford: Oxford University Press.

Kohler-Koch, B. (2010) Civil society and EU democracy: ‘astroturf’ representation? Journal

of European Public Policy 17(1): pp. 100-116

Lambert, R. D., Curtis, J. E., Brown, S.D., Kay, B. J. (1986). Effects of Identification with

Governing Parties on Feelings of Political Efficacy and Trust. Canadian Journals Political

Science. 19 (December) pp. 705-28.

Lijphart, A. (1999) Patterns of Democracy. New Haven/London: Yale University Press

Munck, G.L. and Verkuilen, J. (2002) Conceptualizing and Measuring Democracy: Evaluating

Alternative Indices. Comparative Political Studies 35(1): pp. 5-34

http://www.eiu.com/public/topical_report.aspx?campaignid=DemocracyIndex12

http://www.economics.uni-linz.ac.a/

MYPLACE August 2014


of 431

Nacarino, R. C., De Corte S. and Freudenstein R. (2012) Democracy and Legitimacy in an

Economic Union: Policy Brief. Centre for European Studies

Nevitte, N., and Kanji, M. (2002).Authority Orientations and Political Support: A Cross-

national Analysis of Satisfaction with Governments and Democracy. Comparative Sociology

1 (2002): 387-412. Reprinted in Ronald Inglehart (ed) Human Values and Social Change:

Findings from the World Values Surveys. Leiden-Boston: Brill. 2003.

Norris, P. (1999). Critical Citizens: Global Support for Democratic Governance. Oxford:

Oxford University Press.

Norris, P. (2011) Democratic Deficit: Critical Citizens Revisited. Cambridge University Press:

New York

Pateman, C. (1970). Participation and Democracy Theory. Cambridge: Cambridge University

Press.

Pew Research Center (2013) The New Sick Man of Europe: the European

Unionhttp://www.pewglobal.org/2013/05/13/the-new-sick-man-of-europe-the-european-

union/

Putnam, R. (2000). Bowling Alone: The Collapse and Revival of American Community. New

York: Simon and Schuster.

Putnam, R. (1995). Bowling Alone: America's Declining Social Capital. Journal of Democracy.

6 ( 1): pp. 65-78.

Putnam, R. (1993). Making Democracy Work: Civic Traditions in Modern Italy. Princeton:


Roth, F. Nowak-Lehmann, F. D. and Otter T. (2013), “Crisis and Trust in National and

European Union Institutions – Panel Evidence for the EU, 1999 to 2012”, RSCAS Working

Paper Series 2013/31, European University Institute, Florence, May

http://cadmus.eui.eu/bitstream/handle/1814/26975/RSCAS_2013_31.pdf?sequence=1 RSF

(2014) World Press Freedom Index http://en.rsf.org/

Ruiz-Rufino, R. (2013) Satisfaction with Democracy in Multi-ethnic Countries: The Effect of

Representative Political Institutions on Ethnic Minorities. Political Studies. 61, pp. 101-118.

Sanders, D., Clarke, H., Stewart, M. and Whiteley, P. (2014) Output Oriented Legitimacy:

Individual and System-level Influences on Democracy Satisfaction in Thomassen J. (ed)

Elections and Representative Democracy: Representation and Accountability. (pp. XX-XX)

Oxford: Oxford University Press

Schafer, A. (2013). Affluence, Inequality and Stratification with Democracy. In S. I. Keil and

O. W. Gabriel (Eds.), Society and Democracy in Europe (pp 139-161). London: Routledge.

Schmitter P. C., and Karl T. L. (1991) “What Democracy Is . . . and Is Not,” Journal of

Democracy 2: pp. 76-80

http://www.pewglobal.org/2013/05/13/the-new-sick-man-of-europe-the-european-union/


http://cadmus.eui.eu/bitstream/handle/1814/26975/RSCAS_2013_31.pdf?sequence=1

http://en.rsf.org/

MYPLACE August 2014


of 431

Vatter, A. and Bernauer, J. (2009) The missing dimension of democracy. European Union

Politics 10(3): pp. 335-359

Van der Eijk, C. and Franklin, M. N. (2009) Election and Voters. Palgrave Macmillan: New

York

Weatherford, M. S. (1991).Mapping the Ties that Bind: Legitimacy, Representation and

Alienation. Western Political Quarterly. 44 (June) pp. 251-76.

Weil, F. (1989). The Sources and Structure of Legitimation in Western Democracies: A

Consolidated Madel Tested with Time-Series Data in Six Countries Since World War II.

American Sociological Review. 54 (October) pp. 682-706.

Zmerli, S., Newton, K., Montero, J. R. ( 2007). Trust in People, Confidence in Political

Institutions, and Satisfaction with Democracy. In Jan W. van Deth/José Ramón

Montero/Anders Westholm (eds.). Citizenship and Involvement in European Democracies. A

Comparative Analysis (pp. 35-65). London/New York: Routledge.

MYPLACE August 2014


of 431

Chapter 14: History and Memory

Here we present the cross-country analysis of four questions (Q50, Q51, Q52, Q53) of the

History and Memory block of the MYPLACE survey questionnaire by using merged data sets.

We examine the findings of the 30 data collection locations.

In the first section, we review the literature on collective memory. Then we analyse the

country-level and location-based data of the scales (Q51: Significance of historical events,

Q53: Importance commemorating the past), the Q52: Level of activity in historical

dimensions index and the Q50: Interest in recent history variable that were constructed

based on the four questions. In the last section, we analyse the way and degree of how

respondents’ location, socio-demographic and political characteristics determine their

relation to the three dimensions of history.

It is important to note that of the four questions, the first two (Q51 and Q53) refer to

country history and commemorating the country, while the other two (Q50 and Q52)

involved respondents’ interest in history and history-related activities. Responses to Q51

and Q53 are, obviously determined by respondents’ subjective relation to history and

historical commemoration as well.

According to our hypothesis, all the three variable sets shape respondents’ relation to

memory and history. The impact of location is, determined by the various historical

traditions of European regions; we expect to see a higher impact of education levels and

parental status among socio-demographic variables, while we suppose a determinant effect

of cognitive-type variables (political knowledge, relation to politics) among political

characteristics.

Literature The pioneering French social theorist Maurice Halbwachs (1980; 1992) has argued that

history and memory were contradictory ways of dealing with the past. He valued history

over memory, because he saw history to be a universal scholarship, unlike memory which he

considered to be limited to the lifetime and perspective of a particular community. He saw

history as a reasoned reconstruction of the past rooted in research; a scientific discipline that

tends to be critical and skeptical of human motive and action, and therefore more secular

than what people commonly call memory. In his view history, being more relative, and

contingent on place, chronology, and the scale can be read by or belong to everyone.

His view was both shared and opposed by the theorist Pierre Nora (1984; 1989), who also

saw the two as contradictory, but valued the importance of memory over history. In contrast

MYPLACE August 2014


of 431

to Halbwachs, he argued that while memory is truthful, history is a flawed reconstruction of

the past.

In more recent times, by using the term “collective memory” theorists have challenged the

split between history and memory, seeing them as being complementary (Olick and

Robbins, 1998; Erll, 2005; Till, 2006). The term “collective memory”, however, has been

used as a loosely defined notion when examining issues such as the Holocaust rather than

as a construct that deserves focused attention in its own right (Nora, 1989; Brubaker and

Feischmidt, 2002).

Over the past few decades, collective memory has become a topic of renewed interest in

the humanities and social sciences and is now a key part of emerging interdisciplinary

activity in “memory studies” (Roediger and Wertsch, 2008; Klein, 2000). However, it has

almost as many definitions as investigators writing about it. Perhaps the only generally

agreed-upon feature is that collective memory is a form of memory that transcends

individuals and is shared by a group.

One of the reasons for the problems in defining collective memory is that it is not a topic

that fits neatly within the confines of a single academic discipline. The concept (or

concepts) has been examined by sociologists (e.g., Zerubavel, 2003), anthropologists (e.g.,

Cole, 2001), psychologists (e.g., Middleton and Brown, 2005; Pennebaker et al., 1997),

historians (e.g., Bodnar, 1992), literary analysts (e.g., Young, 1993), and others, but there

has been little contact, let alone coordination, among these efforts. More recently, Yadin

Dudai has argued that “the term ‘collective memory’ actually refers to three entities: a

body of knowledge, an attribute, and a process” (2002: 51). The body of knowledge is a

feature of the culture of the individuals who share some similarity, and individuals may

participate in various different groups (with different collective memories) defined by

generation, countries of origin, locale (e.g., Hungarians), and so on. The attribute is “the

distinctive holistic image of the past in the group” (e.g., World War II veterans in the US

who are referred to as “the greatest generation” by some). The process is the continual

evolution of understanding between the individual and the group, as individuals may

influence and change the collective memory of the group, and the group can change the

individual’s understanding and consciousness of being a member of the group. In Dudai’s

view these three entities capture some (but not all) of the various senses of collective

memory used by scholars in different academic disciplines.

Given the fragmented state of the discussion about collective memory, however, up to the

present day no single, generally accepted definition of the term has emerged. Indeed, the

very term is often used almost interchangeably with others such as “public memory”,

“social memory”, “communicative memory” and “cultural memory” (Assmann, 1992). In

attempting to provide a more precise definition of the term Wertsch and Roediger (2008)

MYPLACE August 2014


of 431

has argued for the need of differentiating between collective memory versus collective

remembering. They contrast collective memory as a static base of knowledge with

collective remembering, which involves the repeated reconstruction of representations of

the past, a process that is often quite contentious. Earlier, Bodnar (1992) saw collective

memory to be more like a space of contestation than a body of knowledge - a space in

which local groups engage in an ongoing struggle against elites and state authorities to

control the understanding of the past. As Reese and Fivus (2008) pointed out, the sites of

the contestation over collective remembering include family discussions museums,

monuments and memorials, history textbooks, and national holidays.

In the Nineteenth Century, Renan (1990; 1882) viewed serious historical research as often

posing a threat to popular efforts at collective remembering. The contradiction between

history and collective memory is central to the argument Nora (1989) makes that ‘‘real

memory’’ has been largely pushed aside, if not eradicated, by the practices of creating

critical historical accounts. He argued that representations of the past based on traditional

practices such as rituals. A similar opposition between history and collective memory can

be found in the account Assmann (1997:23) provided, when he argued that “the past is not

simply ‘received’ by the present. The present is ‘haunted’ by the past and the past is

modelled, invented, reinvented, and reconstructed by the present.’’ (Assmann, 1997:9).

Historians routinely warn against practices of inventing, reinventing, and reconstructing the

past in the service of the present, but this is precisely what is encouraged - indeed

celebrated - in the case of collective remembering. As Novick (1999:3-4) pointed out, the

processes involved are usually tied to schematisation and simplification that stand in

contrast to the aspirations of analytic history.

Some scholars have also drawn a stark opposition between individual and collective

remembering, going so far as to assert that only one or the other exists. Bartlett

(1932;1995), for example, raised the issue of whether collective remembering exists, at

least in the “strong” sense that a group, qua group, “can be usefully characterised as having

some sort of memory in its own right” (Wertsch, 2002:22). In contrast to this, scholars such

as Schudson (1995:347) have questioned whether memory can be anything but collective.

While sharing this view other scholars (Yates, 1966; Bruner, 1990) also argue that

collectives use cultural tools which are items such as written symbols, information storage

in computers, and “mnemotechnics” (Yates, 1966), and what makes collective

remembering collective is that members of a group share the same “cultural tool kit”.

Arguing that memory is “distributed” and relies extensively on semiotic means provided by

cultural, historical, and institutional contexts Hutchins (1995) emphasised that human

agents and the cultural tools they employ must be viewed as integral components of a

memory system.

MYPLACE August 2014


of 431

Contrary to the scholarly study of national identity construction in Europe, which has

always highlighted the essential role played by consensual icons of collective memory for

nation-building on the continent, research on the collective memory of Europeans has

traditionally focused on the conflictual, painful side of memory. The following works are

good examples for this kind of approach:

Ricoeur (2000) La mémoire, l'histoire, l'oubli. Paris: Seuil;

Alexander, et al. (2004): Cultural Trauma and Collective Identity. Berkeley: University of California Press;

Giesen (2004) Triumph and Trauma. Boulder: Paradigm;

Leggewie and Lang (2011). Der Kampf um die europäische Erinnerung (The battle for the European remembrance). Munich: C. H. Beck Verlag;

Pakier and Str th (2013) A European Memory? : Contested Histories and Politics of Remembrance . Berghahn Books

For these authors, collective memory on the European level can hardly be a source of a

positive narrative, as perceptions and interpretations differ widely between the nations

that make up the European patchwork.

Memory in Eastern Europe very often equals trauma because of its experience of 20th

century totalitarianism, genocide, displacement and border changes, while research on

memory requires particular skills and sensitivity. Thus, Eastern Europe is a fertile ground for

memory studies for a number of reasons. This is the only region world-wide to have been

consecutively ruled by Nazism and Communism. Imperial legacies of territorial

dismemberment, belated state-building, and massive forced migration were compounded

by the unique material devastation and disastrous psychological consequences of World

War II. The destruction of social elites, urban spaces, and ethnic groups caused divided

memories of victimhood but also desires for vengeance and retribution. The diversity of

national self-images, cultural policies, and international relations across Eastern Europe can

be largely explained by different beliefs about historic events involving large-scale human

suffering. The public memory of these Twentieth-Century traumas mediates the variety of

ways in which nations develop in the post-socialist space.

Unlike Western Europe, where the initial post-World War II repression of memory has been

transformed into political recognition of victims and has contributed to the politics of

regret, much of Eastern Europe is in the grip of a ‘presence of the past’. After the

breakdown of communism, post-socialist societies such as Russia have launched new

attempts at ‘organised forgetting’ by constructing patriotic master narratives and silencing

the complexity of the past. The memories of Twentieth-Century tragedies travel onto

political agendas by mass media, literature, film, the visual arts, and new Internet media.

More recently several researchers made an attempt to explore this historically unique

phenomenon:

MYPLACE August 2014


of 431

Mink and Neumayer (eds.) (2013) History, Memory and Politics in Central and Eastern Europe : Memory Games . Palgrave Macmillan;

Blacker et al. (2013) Memory and Theory in Eastern Europe. Palgrave Macmillan;

Rutten et al. (2013) Memory, Conflict and Social Media : Web Wars in Post-Socialist States . Routledge;

Kattago (2013) Memory and Representation in Contemporary Europe : The Persistence of the Past , Ashgate Publishing, Ltd.

Maggia et al. (2010) Contemporary Photography from Eastern Europe : History, Memory, Identity. Skira

14.1 Interest in recent history (Q50)

The distribution of responses for Q50 (Would you say that you are interested in the recent

history (last 100 years) of [COUNTRY]?) is the following regarding the entire sample (N=

16.935): not at all: 14.3 %, a little: 55.8 %, very interested: 29.0 %, missing:1.8%

MYPLACE August 2014


of 431

Q50: Interest in recent history


The rate of interest in country history in the entire sample (n= 16780): not at

all: 14.4 %, a little: 56.3%, very interested: 29.3%. The highest mean value of

interest among countries is: not at all: 1.6 %, a little: 44.4%, very interested:

54.0% (eastern Germany), while the lowest mean value of interest is: not at

all: 27.0 %, a little: 64.1 %, very interested: 8.9 % (Slovakia). The highest

mean value of interest among locations is: not at all: 1.8 %, a little: 41..4%,

very interested: 56.8% (Jena, eastern Germany), while the lowest mean

value of interest is: not at all: 44.1 %, a little: 50.3%, very interested: 5.6%

(Vyborg, Russia)

Graph 14.1 : Percentage of Interest in recent history variable by

location

Narrative

There are clear clusters of locations; with locations in Georgia, Greece and

both eastern and western Germany all in the top third, whose respondents

consider the impact of historical events the most important. Hungary,

Slovakia and Latvia are, clustered in the bottom third.

Several countries had similar distributions of two locations; notable

exceptions with significant differences among locations in the following

cuntries were: Estonia (Tartu and Narva) p ≤0.001, Finland (Kuopio and

Lieksa and Nurmes) p ≤0.05, Germany- (Bremen and Bremenhaven) p ≤0.05,

Hungary (Sopron and Ózd) p ≤0.001, Latvia (Agenskalns and Jauna Forstadte)

p ≤0.001, Portugal (Lumiar and Barreiro) p ≤0.001, Russia (Kupichno and

Vyborg) p ≤0.001, Spain (Vic and Sant Cugat del Vallés)p ≤0.001

MYPLACE August 2014


of 431

14.2 Significance of historical events (Q51)

Respondents’ views on the significance of historical events for the country were gathered by

asking the following nine events. Respondents rated each on a five point scale ranging from

‘not important at all’ to ‘very important. (the mean values for items: Q51_1. World War I

(1914 - 1918) mean: 2.94; Q51_2. World War II (1939 - 1945) mean: 3.21; Q51_3. Cold War

(1945 - 1989) mean: 2.58; Q51_4. Fall of the Berlin Wall and the end of socialist regimes

(1989-1991) mean: 2.68; Q51_5. The Holocaust, mean: 2.44; Q51_6. Fascism, mean: 2.72;

Q51_7. Communism, mean: 2.76; Q51_8 The NY terrorist attacks on September 9th 2001.,

mean: 2.07; Q51_9. Membership in the European Union) mean: 3.18.)

Based on the order of mean values, respondents consider World War II and membership in

the European Union the most important from their countries’ perspectives among the listed

events and periods, and NY terrorist attacks September 9th 2011 and the Holocaust the least

important.

The results of the factor analysis (Extraction Method: Maximum Likelihood, Rotation

Method: Direct Oblimin) were two factors. If more than one factor is derived and the

correlation among those factors is less than .32, we have used varimax rotation for the

analysis. We excluded two variables (Q5.8, Q5.9) from the next factor analysis as

communality was less than 0.2.

Seven items load under one factor explaining 44.5 % of the total variance. These seven

items constitute a good scale ranging from zero to 28. Higher score in the scale indicates

greater significance of historical events. (Mean: 19.48, std. dev. = 5.045, n=14214).

MYPLACE August 2014


of 431

Q51: Significance of historical events


The mean value of the impact of recent historical events in the full sample

(n= 14214, sd=5.045) is 19.48. Among countries, the highest mean value is

23.20 (eastern Germany) and the lowest mean value is 17.55 (Estonia).

Among locations, the highest mean value is 23.64 (Jena, eastern Germany)

and the lowest mean value is 15.00 (Narva area, Estonia).

Graph 14.2 : Mean significant of historical events scale by location

Narrative

There are clear clusters of locations; with locations in Russia and both

eastern and western Germany all in the top third, with the indication of the

highest importance of recent historical events. Denmark, Croatia, Georgia

and Portugal are, clustered in the bottom third.

Most countries had similar mean score between the two locations; notable

exceptions were; primarily and substantially Estonia (Tartu and Narva) 4.85,

Slovakia (Trnava and Rimavska Sobota) 1.76, Finland (Kuopio and Lieksa and

Nurmes) 1.36, Greece (Argyroupouli and New Philadelphia ) 1.01.

MYPLACE August 2014


of 431

14.3 Interest in history index (Q52)

In order to measure how interested the respondents were in history, they were asked how

frequently they did the following 12 things in the last 12 months.

The distribution of responses for the 12 items of Q52 is the following regarding the total

sample:

The distribution of responses per item (%), n: never, o: occasionally, r: regularly. Q52_1:

Played a video game that re-enacts events in the 20th century (n:67.1 o:24.9 r:8.1 ), Q52_2:

Watched a film that was set in recent historical circumstances(n: 22.8, o:58.2, r: 19.0 ),

Q52_3: Watched a documentary about events in the 20th century (n:28.7, o:55.9, r:15.3 ),

Q52_4: Read a novel set in recent historical circumstances (n: 60.8, o:32.3, r:6.9), Q52_5:

Read a non-fiction publication about the recent past (n: 62.3, o: 30.9, r: 6.8), Q52_6: Visited

a museum with an exhibition about the recent past (n: 56.4, o: 39.1, r:4.5), Q52_7:

Participated in a discussion about history at school or in college (n:54.0, o:34.0, r:12.0 ),

Q52_8: Talked with your parents about the past (n: 27.0, o: 57.4, r:15.6), Q52_9: Talked with

your grandparents about the past (n: 41.7, o: 44.1 r:14.2 ), Q52_10: Attended a veterans’

parade or showed support during a remembrance day (n: 75.9, o:18.5, r:5.6), Q52_11:

Researched family history (n: 54.2, o:38.0, r: 7.9), Q52_12: Visited a webpage with historic

content / discussions( n:54.6, o: 36.4, r:9.0 )

Based on the responses on the 12 listed activities, youths performed the following activities

the most frequently over the last 12 months: 1. Watched a film that was set in recent

historical circumstances, 2. Talked with their parents about the past, 3. Watched a

documentary about events in the 20th century. The least frequently performed activities

involved the following: 1. Attended a veterans’ parade or showed support during a

remembrance day, 2. Read a novel set in recent historical circumstances, 3. Played a video

game that re-enacts events in the 20th century, 4. Read a non-fiction publication about the

recent past.

The fact that more than one-fifth (22.8%) of the respondents has not even watched films

that were set in recent historical circumstances, which is the most frequent activity among

them, implies the low intensity of past-related actions among youths. Another characteristic

of the order set by respondents is that for youths, films are the most likely and books are

the least likely resources for learning past events, although we have to emphasise the

relatively frequent communication with parents.

We constructed an index based on the 12 variables related to these questions.

MYPLACE August 2014


of 431

Originally, responses were recorded under one of the three categories: never, occasionally,

regularly. In constructing the index, respondents who reported ‘occasionally’ or ‘regularly’

for each aspect was counted. The index ranges from 0-12 in which highest score indicates

greater interest in history (Mean=5.87, Std.Dev: 3.172,n=15236).

MYPLACE August 2014


of 431

Q52 Interest in history index


The mean value of the importance of recent historical events’ impact

in the entire sample (n= 14214, sd=5.045) is 19.48. Among countries,

the highest mean value is 7.41 (eastern Germany) and the lowest

mean value is 3.72 (Georgia). Among locations, the highest mean

value is 7.46 (Rostock, eastern Germany) and the lowest mean value is

3.70 (Kutaisi, Georgia).

Graph 14.3 : Mean Interest in history index by location

Narrative

There are clear clusters of locations; with locations in Russia and both

eastern and western Germany all in the top third, with the indication

of the highest importance of recent historical events. Denmark,

Croatia, Georgia and Portugal are, clustered in the bottom third.


notable exceptions were; primarily and substantially Estonia (Tartu

and Narva) 4.85, Slovakia (Trnava and Rimavska Sobota) 1.76, Finland

(Kuopio and Lieksa and Nurmes) 1.36, Greece (Argyroupouli and New

Philadelphia ) 1.01.

MYPLACE August 2014


14.4 Importance commemorating the past (Q53)

According to the distribution of valid responses for Q53 (Please indicate on a scale of 0 to 10

how important it is for [COUNTRY] to commemorate its past where 0 indicates ‘not

important at all’ and 10 ‘very important’?) we can state that only 7.7% of respondents think

that it is not important (rate 0-4) to, 17.4% think that it is maybe important (rate 5) to while

two-third (74.9%) think that it is important to commemorate their past in their countries.

Table 14.4: Please indicate on a scale of 0 to 10 how important it is for [COUNTRY] to

commemorate its past where 0 indicates ‘not important at all’ and 10 ‘very important’?

(percentage)

Frequency Percent Valid

Percent

Valid

0: not at all important 148 0.9 0.9

1 103 0.6 0.6

2: 210 1.2 1.3

3: 372 2.2 2.2

4: 454 2.7 2.7

5: 1615 9.5 9.7

6: 1383 8.2 8.3

7: 2420 14.3 14.5

8: 3226 19.0 19.3

9: 2307 13.6 13.8

10: very important 4456 26.3 26.7

Total 16694 98.6 100

Missing

interviewer error 1 0.0

question not asked 2 0.0

refusal 57 0.3

don't know 178 1.1

system-missing 3 0.0

Total 241 1.4

Total 16935 100.0

MYPLACE August 2014


Q53: Importance commemorating the past


The mean value of the importance of commemorating the past in the entire

sample (n= 16694, sd=2.180) is 7.66. The highest mean value of the

importance of commemorating the past among countries is 9.47 (Georgia)

and the lowest mean value is 6.82 (Latvia). The highest mean value of the

importance of commemorating the past among locations is 9.50 (Telavi,

Georgia) and the lowest mean value is 6.14 (Jauna & Forstadte, Latvia).

Graph 14.4 : Mean importance commemorating the past variable

by location

Narrative

There are clear clusters of locations; with locations in Georgia and Greece in

the top third, with the implication of the highest importance of

commemorating the past. Denmark, Hungary, Slovakia and Spain are,



exceptions were; primarily Latvia : 1.36 (Agenskalns- Jauna Forstadte)

Estonia: 1.55( Tartu- Narva area), Russia 1.13 (Kupchino-Vyborg).

MYPLACE August 2014


14.5 Multivariate analysis

The relationship of the ordinal-level Q50 variable (Interest in recent history variable) and the

other three interval-level variables (Q53: Importance commemorating the past variable,

Q51: Significance of historical events scale, Q52: Level of activity in historical dimensions

index) in the History and Memory topic is inevitable, the more respondents are interested in

recent history, the more important they consider it to commemorate historical events, the

more frequently they perform history-related activities and the higher impact they imply for

these on country history. Relationships are significant statistically (Independent-Samples

Kruskal-Wallis Test): (1. Q50 - Q51 scale (Test stat.: 807.037, df: 2, Sig. (2-sided test): 0.000),

2. Q50 - Q52 index (Test stat.: 1781.841, df: 2, Sig. (2-sided test): 0.000), 3. Q50 - Q53

variable (Test stat.: 1752.738, df: 2, Sig. (2-sided test): 0.000).

Table 14.5.1 The relation of variable Q50 with the Q51, Q53 scales and the Q52 index

(means)

Interest in recent history (Q50) Q51 Q52 Q53

not at all 17.04 3.68 6.24

a little 19.19 5.79 7.59

very interested 21.05 7.19 8.50

The further examination of the three variables (Q51 scale, Q52 index, Q53 variable) –

besides the strong relation among them – is justified by the fact that with the application of

the same method, the outcomes of the multi-variable analyses are comparative.

In the following section, we study the relation of respondents’ relation to, history and

historical memory, the degree to which respondents’ perception of historical past and

history-related activities are, determined by their locations, socio-demographic and political

characteristics.

The dependent variables of the multiple linear regression-ordinary least squares, enter

method are the following: Model 1: Q51: Significance of historical events scale, Model2:

Q52: Level of activity in historical dimensions index, Model3: Q53: Importance

commemorating the past variable. The independent variables for all the three models are

constituted by the same variable set: 1. Locations, 2. Socio- demographic characteristics, 3.

Political characteristics. The description of the independent variables is the following:

1. Location variables. We created 29 dummy variables based on the 30 categories of the

Locations variable. The variable with the smallest mean and standard deviation – as

compared to the standard deviation of the dependent variable – serves as the reference

category.

MYPLACE August 2014


2. Socio-demographic variables: Age (number of years), Gender (1: male, 0: female),

Parental social class (scores on social class from Q73, Q77,Q76 and Q80), Identity (1:

majority, 0: minority), Employment status1 (1: education, 0: not), Employment status2

(1:employed, 0: not), Employment status3 (1: unemployed, 0: not), Religion (1: religious, 0:

not religious).

3. Political characteristics variables

3.1 Satisfaction with democracy (Q42). The percentage distribution of responses for Q42

(On the whole, how satisfied are you with the way democracy works in [country]?) is the

following regarding the total sample (N= 16.935): 0 (extremely dissatisfied): 6.0 %, 1: 3.6 %,

2: 6.4 %, 3: 9.5 %, 4: 9.9 %, 5:19.9 %, 6: 12.2 %, 7: 13.8 %, 8:9.7 %, 9: 3.6 %, 10: 2.0 %,

missing: 3.2 %, mean: 5.01, std.dev: 2.438.

3.2 Left-right scale (Q21). The percentage distribution of responses for Q21 (On a scale of 0

to 10, where 0 is left and 10 is right, would you say that you personally are left or right-

wing?) is the following regarding the total sample (N= 16.935): 0 (left): 3.7 %, 1: 2.6 %, 2:

6.0 %, 3: 8.8 %, 4: 8.1 %, 5: 25.4 %, 6: 6.6 %, 7: 6.2 %, 8: 5.0 %, 9: 2.3 %, 10: 3.4 %, missing:

21.7 %, (do not know: 16.5) mean: 4.84, std.dev: 2.344.

3.3 Political knowledge (Q28S). We constructed the variable based on the responses of

questions Q28.1-Q28.3. The values of the constructed scale range from 0 (all wrong or DK)

to 3 (all correct).

3.4 Interest in politics (Q2). Question Q2 referred to the interest in politics among

respondents’ direct personal environment (father, mother, brother or sister, grandparent,

boyfriend/ girlfriend/partner, best friend). Following seven items, each measuring degree of

interest (from ‘not at all interested’ to ‘very interested’) scored in four points.

The factor analysis resulted in only one factor, so we used principal component analysis. All

items load under one factor (principal component) explaining almost 47.43 % of the total

variance. These seven items constitute a good scale (Cronbach Alpha: 0.812) ranging from

zero to 21. Higher score in the scale indicates more frequent discussion on political issues

with primary group members.

3.5 Effectiveness of the "soft" action scale ( Q19/9), Effectiveness of "hard" action scale (

Q19/2). Question Q19 involved the perception of 11 activities, that is, to what degree do

these activities influence politics according to respondents. In the factor analysis, we have

first used oblimin rotation. The result of the factor analysis were, two factors derived and

the correlation among those factors is -0.371, so we have used varimax rotation for the

analysis. The two factors are derived from the following items: factor 1 (9 items) and factor

MYPLACE August 2014


2 (6 items). These 11 items load under 2 factors: intensive activity, including violent and

illegal protest activities under factor 2 and the remaining 9 items (less intense and

legitimate) under factor 1. The characteristics of Factor 1 are the following: from a statistical

analysis point of view, these scales under one factor/latent structure work well as they have

excellent level of reliability (Cronbach alpha:0.884). These 9 items can be described

meaningfully under the concept of ‘soft action’: protests, which may be legal. Therefore,

they could be, used as a scale and named as ‘Effectiveness of the "soft" action scale’. The

characteristics of Factor 2 are similar (Cronbach alpha:0.852). It can be named as

‘Effectiveness of the "hard" action scale’. Effectiveness of "soft" action scale ranges from 0

to 90, with higher score indicating greater effectiveness of "soft" action to influence politics.

Effectiveness of "hard" action scale ranges from 0 to 20. Higher score in the scale indicates

greater effectiveness of "hard" action to influence politics.

Table 14.5.2: Multiple linear regression-ordinary least squares Model1 - Model3 (enter

method, cell data include: 1. Standardised Coefficient (Beta), 2. T value, 3. Sig. : * p < 0.05,

**p < 0.001)


Model1

dependent variable:

Q51 scale

Model1

dependent variable:

Q52 index

Model3

dependent variable:

Q53 variable

Podsljeme (CRO) -.078 (-3.967) ** .007 (.424) .005 (.255)

Pescenica Zitnjak (CRO) -.085 (-4.202) ** Referent category .022 (1.110)

Odense East (DEN) -.067(-3.554) ** -.020 (-1.170) -.075 (-4.158) **

Odense Center (DEN) -.081 (-4.032) ** -.008 (-.433) -.097 (-5.081) **

Narva area (EST) -.083 (-4.417) ** -.003 (-.162) -.012 (-.701)

Tartu (EST) -.006 (-.272) .030 (1.639) .082 (4.053) **

Lieksa and Nurmes (FIN) -.030 (-1.824) .004 (.272) .008( .491)

Kuopio (FIN) -.053 (-2.894) * .040 (2.437) * -.007 (-.382)

Kutaisi (GEO) -.056 (-3.397) * -.131 (-8.345) ** .104 (6.141) **

Telavi (GEO) -.043 (-2.627) * -.101 (-6.565) ** .100 (6.109) **

Bremen (GER-W) .076 (3.778) ** -.003 (-.165) -.014 (-.731)

Bremerhaven (GER-W) .020 (1.111) .009 (.537) -.041 (-2.422) *

Jena (GER-E) .136 (5.859) ** .021 (1.050) -.020 (-.946)

Rostock (GER-E) .124 (5.296) ** .041 (2.011) * -.049 (-2.245) *

New Philadelphia (GRE) Reference category -.083 (-4.729) ** .085 (4.414) **

Argyroupouli (GRE) .021 (1.123) -.105 (-5.966) ** .046 (2.380) *

Sopron (HUN) .061 (2.898) * -.180 (-9.868) ** -.077 (-3.754) **

Ózd (HUN) .024 (1.334) -.086 (-5.182) ** -.008 (-.442)

Agenskalns (LAV) .013 (.607) .089 (4.702) ** .014 (.723)

Jauna Forstadte (LAV) .043 (2.029) * .034 (1.802) -.073 (-3.842) **

Lumiar (POR) -.061 (-3.176) * -.037 (-2.198) * -.066 (-3.521) **

Barreiro (POR) -.099 (-5.319) ** -.140 (-8.338) ** -.014 (-.762)

Kupchino (RUS) .039 (2.104) * .048 (2.885) * .067 (3.641) **

Vyborg (RUS) .153 (7.216) ** -.088 (-4.729) ** -.083 (-3.985) **

Rimavskva Sobota (SLO) -.039 (-1.977) * -.064 (-3.628) ** -.056 (-2.943) *

MYPLACE August 2014


Tranava (SLO) -.044 (-1.945) -.140 (-7.059) ** -.055 (-2.454) *

Vic (SPA) -.037 (-1.990) * -.025 (-1.450) -.042 (-2.284) *

Sant Cugat del Vallés (SPA) -.063 (-3.198) * -.024 (-1.352) -.087 (-4.568) **

Coventry (UK) .030 (1.504) .066 (3.583) ** Reference category

Nuneaton (UK) .028 (1.550) .078 (4.657) ** .044 (2.509) *

Age .022 (1.397) -.115 (-7.766) ** .026 (1.658)

Gender -.030 (-2.206) * .056 (4.330) ** -.005 (-.390)

Parental social class .060 (4.058) ** .063 (4.544) ** .040 (2.768) *

Identity .069 (4.198) ** .017 (1.070) .051 (3.154) *

Employment status1 .084 (3.092) * .045 (1.730) .037 (1.401)

Employment status2 .006 (.217) .031 (1.290) .000 (-.002)

Employment status3 -.011 (-.558) .006 (.304) .015 (.779)

Religion .027 (1.824) -.010 (-.700) .046 (3.090) *

Satisfaction with democracy .030 (1.909) -.067 (-4.499) ** .080 (5.095) **

Left-right scale -.044 (-2.988) * .034 (2.429) * .047 (3.253) *

Political knowledge .120 (7.647) ** .123 (8.215) ** .075 (4.803) **

Interest in politics .078 (5.167) ** .183 (12.834) ** .050 (3.326) *

Effectiveness of "soft" action .051 (3.057) * .110 (6.921) ** .204 (12.327) **

Effectiveness of "hard" action .046 (2.971) * .035 (2.371) * -.104 (-6.793) **

Adjusted R Square 0.215 0.256 0.170

The outcomes of the regression models:

Model 1

The independent variables applied explain 21.5% of the influence of recent historical events

on respondents’ country history (Q51).

The perception of the impact of historical events was not, significantly influenced if

respondents live in either of the two UK locations (Coventry, Nuneaton) or any of the

following locations: Tartu (EST), Lieksa and Nurmes (FIN), Bremerhaven (GER-W),

Agenskalns (LAV).

Six locations are, characterised by significant beta values in this model. Consequently,

respondents indicated a greater impact of historical events if they live in either of the two

eastern German locations (Jena, Rostock, GER-E) or any of the following locations: Bremen

(GER-W), Vyborg (RUS), Sopron (HUN), Jauna Forstadte (LAV).

The effect is still significant related to the other locations as well but negative beta values

imply that respondents living in any of the remaining 14 locations implied a smaller level of

influence for historical events.

Among socio-demographic variables, 4 had significant effect: respondents rated higher

importance for historical events if they are women, majority, currently studying or have

parents with higher social class.

MYPLACE August 2014


Of the third set of independent variables, political knowledge had the greatest influence

with the highest beta value among significant variables. By taking into consideration the

other significant variables we can state that the higher level of political knowledge and

interest in politics they have, the more they think that politics can be influenced with

different activities, respondents implied a higher level of influence of historical events.

Respondents’ political self-assessment also, impacts, significantly the perception of

historical events: being left-wing involves a greater likelihood for implying historical events’

greater impact on country history.

Model 2

The independent variables applied explain 25.6% of the influence of respondents’ frequency

for performing past-related activities and thus their interest in history (Q52).

Interest in history, that is, the frequency of performing past-related activities are not

influenced significantly if respondents live in either of the two Danish (Odense East, Odense

Center), Estonian (Narva area, Tartu), Spanish (Vic, Sant Cugat del Vallés) or western-

German (Bremen, Bremenhaven) locations or any of the following locations: Podsljeme

(CRO), Lieksa and Nurmes (FIN), Jena (GER-E), Jauna Forstadte (LAV).

Six locations are, characterised by significant positive beta values in the model.

Consequently, performing past-related activities is more likely if respondents live in either of

the UK locations (Coventry, Nuneaton) or any of the following: Kuopio (FIN), Rostock (GER-

E), Agenskalns (LAV), Kupchino (RUS).

The effect is still significantly related to the other locations as well but negative beta values


influence for historical events.


frequency for performing past-related activities if they are male, younger or have parents

with higher social class.

Of the third set of independent variables, interest in politics had the greatest influence with

the highest beta value among significant variables. By taking into consideration the other

significant variables we can state that the higher level of political knowledge and interest in

politics they have, the more they think that politics can be, influenced with different

activities (mostly soft political actions); respondents implied a higher frequency for

performing past-related activities. Respondents’ political self-assessment and satisfaction

with democracy also, impact significantly the frequency of performing past-related

MYPLACE August 2014


activities: being right-wing or more satisfied with democracy in the given country involve a

greater likelihood for implying more frequently performed past-related activities.

Model 3

The independent variables applied explain 17.0% of the importance of commemorating the

past (Q53).

The importance of commemorating the past is not influenced significantly if respondents

live in either of the two Finnish (Lieksa and Nurmes, Kuopio), Croatian (Podsljeme, Pescenica

Zitnjak) locations or any of the following locations: Narva area (EST), Bremen (GER-W), Jena

(GER-E), Ózd (HUN), Agenskalns (LAV), Barreiro (POR) .

Seven locations are, characterised by significant positive beta values in the model.

Consequently, respondents indicate a higher level of importance for commemorating the

past if they live in either of the Georgian (Kutaisi, Telavi, GEO) or Greek (New Philadelphia,

Argyroupouli, GRE) locations or any of the following: Tartu (EST), Kupchino (RUS), Nuneaton

(UK).

The effect is still significantly related to the other locations as well but negative beta values


importance for commemorating the past.


importance for commemorating the past if they are the majority, religious or have parents

with higher social class.

Of the third set of independent variables, all the three political characteristic variables have

a significant effect. The support of influencing politics with soft political actions had the

strongest significant effect as the Effectiveness of “soft” scale had the highest beta value. By

taking into consideration the other significant variables, we can state that the higher level of

political knowledge and interest in politics they have, the more they think that politics can

be influenced soft political activities, respondents implied higher importance for

commemorating the past. Respondents’ political self-assessment and satisfaction with

democracy also, impact significantly the importance of commemorating the past: being left-

wing or more satisfied with democracy in the given country involve a greater likelihood for

implying higher importance for commemorating the past.

MYPLACE August 2014


14.6 Summary


and 56% a little interested . Respondents consider World War II and membership in the

European Union the most important from their countries’ perspectives among the listed

events and periods, and NY terrorist attacks September 9th 2011 and the Holocaust the

least important.










Our hypotheses concerning the independent variables detailed in the introductory section

are verified by the outcomes of regression models.

The locations, socio-demographic and political characteristics variable’s impact politics-

related activities the most (22%) and the importance for commemorating the past the least

(17%).

Among locations, the Lieksa/Nurmes Finnish location did not have significant effect in any

of the three models. The impact of Tartu (EST) and Bremerhaven (GER-W) was not

significant in the first two models, and the same applies to the Narva area (EST), Bremen

(GER-W), Podsljeme (CRO) and Jena (GER-E) with regard to Model 2 and Model 3, while

Agensalns (LAV) was not significant in Model 1 and Model 3. We can state that the

perception of history and the past is, influenced the least in the Finnish, Estonian and

western German locations (non significant effect).

Three locations had significant roles: Rostock (GER-E) had significant positive beta values in

Model 1 and Model 2 and the same applies to Nuneaton (UK) and Kupchino (RUS) with

regard to Model 2 and Model 3. We can conclude that we are more likely to find a higher

level of importance attached to historical events and commemorating the past if

respondents live in any of these three locations.

According to our results, two socio-demographic variables determine respondents’ relation

to the historical past. Parental social class and belonging to the majority society significantly

MYPLACE August 2014


increase the importance of historical events and commemorating the past. The sex variable

had a significant effect in Model 1 and Model 2 but women tend to attache importance to

historical events while men favour history-related activities.

The perception of the importance of history, the past and historical

memory/commemoration is clearly influenced by political relations; the level of political

knowledge, interest in politics and the support of mainly soft political actions as a means for

influencing politics had significant effects in all three models. A major finding is that past-

related activities participated in characterise those the most who are dissatisfied with

democracy and are right-wing. In contrast being left-wing increases the likelihood of

attaching a higher level of importance to history and commemoration.

References

Assmann, J. (2008) Of God and Gods: Egypt, Israel, and the Rise of Monotheism (2008)

The University of Wisconsin Press.

Assmann, J. (2007). Cultural memories and national narratives: With some relation to the

case of Georgia. White paper report prepared for the Georgian Ministry of Education.

Based on the working meeting ‘‘Negotiating a new national narrative in Georgia’’.

Caucasus Context, 3(1), 40-41.

Assmann, J. (1992) Das kulturelle Gedächtnis. Schrift, Erinnerung und politische Identität

in frühen Hochkulturen. Munich.

Bartlett, F. C. (1995). Remembering: A study in experimental and social psychology.

Cambridge, UK: Cambridge University Press. (Original work published 1932)

Bodnar, J. (1992). Remaking America: Public Memory, Commemoration, and Patriotism in

the Twentieth Century. Princeton, NJ: Princeton University Press

Brubaker, R. and Feischmidt, M. (2002): 1848 in 1998: The Politics of Commemoration in

Hungary, Romania, and Slovakia: Comparative Studies in Society and History, Vol. 44, No.

4 (Oct.,), pp. 700-744.

Ciobanu, M. (2009). Criminalising the Past and Reconstructing Collective Memory: The

Romanian Truth Commission. Europe-Asia Studies, 61, 313-336.

Cole, J. (2001). Forget colonialism? Sacrifice and the art of memory in Madagascar.

Berkeley, CA: University of California Press.

Dudai, Y. (2002). Memory from A to Z: Keywords, concepts and beyond. Oxford, UK:


MYPLACE August 2014


Erll, A., (2005). Kollektives Gedächtnis und Erinnerungskulturen (Stuttgart and Weimar)

Fine, G. (2001). Difficult Reputations: Collective Memories of the Evil, Inept and

Controversial. Chicago: University of Chicago Press.

Halbwachs, M. (1980). The collective memory [translated by Francis J. Didder, Jr. and

Vida Yazdi Ditter]. New York: Harper and Row.

Halbwachs, M. (1992). On collective memory [edited, translated, and with an

introduction by Lewis A. Coser]. Chicago. University of Chicago Press.

Hutchins, E. (1995). How a cockpit remembers its speed. Cognitive Science, 19, 265-288.

Klein, K. L. (2000).`On the emergence of 'memory' in historical discourse’ Representations

69.

Middleton, D., and S. D. Brown. (2005). The social psychology of experience: Studies in

remembering and forgetting. San Francisco, CA: Sage.

Nora, P. (1989). Between memory and history: Les lieux de me´ moire. Representations,

26, 7-25.

Novick, P. (1999). The Holocaust in American life. Boston: Houghton Mifflin Company.

Olick, J. (2007). The Politics of Regret: On Collective Memory and Historical Responsibility.

London: Routledge.

Olick, J. K. and Robbins J., (1998); `Social memory studies. From 'collective memory' to

the historical sociology of mnemonic practices’, Annual review of sociology 24.

Pennebaker, J. W., Paez, D., and Rime´ , J. (Eds.). (1997). Collective memory of political

events: Social psychological perspectives. Mahwah, NJ: Lawrence Erlbaum Associates

Inc.

Reese, E., and Fivush, R. (2008). The development of collective remembering. Memory,

16, 201-212.

Renan, E. (1990). What is a nation? In H. K. Bhabha (Ed.), Nation and narration. London:

Routledge. (Original work published 1882)

Ricoeur, P. (2000). La mémoire, l'histoire, l'oubli. Paris: Seuil.

Roediger, H. L., and Wertsch, J. V. (2008). Creating a new discipline of memory studies.

Memory Studies 1(1), 5-17.

Rowe, S., Wertsch, J., and Kosyaeva, T. (2002). Linking Little Narratives to Big Ones:

Narrative and Public Memory in History Museums. Culture and Psychology, 8, 96-112.

Schudson, M. (1995). Dynamics of distortion in collective memory. In D. Schacter (Ed.),

Memory distortion: How minds, brains, and societies reconstruct the past (pp. 346-364).

Cambridge, MA: Harvard University Press.

Till, K. E., `(2006) Memory studies,` History workshop journal 62.

MYPLACE August 2014


Wertsch, J. (2002). Voices of Collective Remembering. Cambridge: Cambridge University

Press.

Wertsch, J. (2007). Collective Memory. In J. Valsiner and A. Rosa (Eds.) The Cambridge

Handbook of Sociocultural Psychology. Cambridge: Cambridge University Press, pp. 645-

660.

Wertsch, J. (2008). Collective Memory and Narrative Templates, Social Research, 75, 133-

156.

Yates, F. (1966). The art of memory. Chicago: University of Chicago Press.

Young, J. (1993). The texture of memory: Holocaust memorials and meaning. New Haven,

CT: Yale University Press.

Zerubavel, E. (2003). Time maps: Collective memory and the social shape of the past.

Chicago: University of Chicago Press.

MYPLACE August 2014


Chapter 15: European Issues

This chapter examines questions, which specifically relate to Europe. These questions are

also analysed in other chapters, and individually, or as part of a constructed index.

Respondents’ attitudes to European Issues were measured using a range of questions,

including respondents’ level of interest in issues to do with Europe and their level of interest

in the European Union. Other questions explored the level of trust in the European

Commission and the historical importance of having membership in the European Union.

MYPLACE August 2014


Summary of Results

Table 15: Results of the five questions in the MYPLACE questionnaire, which deal with issues

to do with Europe and the European Union.

Location Country Q

1_5

: Lev

el o

f in

tere

st in

issu

es t

o d

o w

ith

Eu

rop

e (

1)

Q5

_6: L

evel

of

Inte

rest

in t

he

Euro

pea

n U

nio

n (

2)

Q7

_7: L

evel

of

tru

st in

th

e

Euro

pea

n C

om

mis

sio

n (

3)

Q2

6_5

: Me

mb

ersh

ip o

f th

e

Euro

pea

n U

nio

n g

reat

ly

ben

efit

s th

is c

ou

ntr

y (4

)

Q5

1_9

: His

tori

cal i

mp

ort

ance

of

the

cou

ntr

y h

avin

g

me

mb

ersh

ip in

th

e Eu

rop

ean

Un

ion

(5

)

Podsljeme Croatia 51.3% 5.48 4.73 35.8% n/a

Pescenica Croatia 54.5% 5.32 4.28 30.7% n/a

Odense East Denmark 58.8% 4.46 5.77 56.8% 90.1%

Odense Center Denmark 65.4% 4.84 5.89 57.3% 93.5%

Narva area Estonia 63.0% 4.93 5.24 45.7% 83.7%

Tartu Estonia 64.9% 5.62 6.04 49.3% 91.5%

Lieksa & Nurmes Finland 45.7% 4.14 5.46 41.3% 73.3%

Kuopio Finland 64.6% 5.41 6.26 46.2% 91.2%

Kutaisi Georgia 54.9% 7.10 6.12 n/a n/a

Telavi Georgia 57.1% 7.54 6.60 n/a n/a

Bremen Western Germany 70.7% 6.09 5.62 73.0% 90.0%

Bremerhaven Western Germany 59.9% 5.70 5.48 66.6% 85.2%

Jena Eastern Germany 72.4% 5.83 5.79 70.6% 91.1%

Rostock Eastern Germany 71.9% 5.85 5.69 69.7% 87.8%

New Philadelphia Greece 59.3% 5.89 3.66 27.9% 89.9%

Argyroupouli Greece 62.9% 5.89 3.45 39.1% 88.6%

Sopron Hungary 55.0% 4.58 4.69 39.0% 81.6%

Ozd Hungary 60.4% 5.66 4.72 26.0% 77.0%

Agenskalns Latvia 59.7% 5.61 5.45 43.1% 88.1%

Forstate & Jaunbuve Latvia 40.1% 4.41 4.59 40.6% 87.2%

Lumiar Portugal 83.7% 6.65 5.39 60.3% 86.5%

Barreiro Portugal 78.1% 6.51 5.50 59.0% 85.2%

Kupchino Russia 51.7% 3.95 4.26 n/a n/a

Vyborg Russia 37.9% 3.97 4.96 n/a n/a

Rimavska Sobota Slovakia 36.0% 4.34 4.76 25.8% 57.5%

Trnava Slovakia 35.6% 4.32 4.55 43.8% 70.1%

Vic Spain 64.0% 6.34 5.35 66.1% 94.9%

Sant Cugat Spain 67.5% 6.48 5.47 67.9% 97.1%

Coventry UK 68.6% 5.77 5.62 54.3% 77.9%

Nuneaton UK 41.2% 4.23 4.66 46.6% 64.1%

MYPLACE August 2014


1 = proportion of respondents stating ‘very interested’ or ‘interested’

2 & 3 = mean value for each location

4 = proportion of respondents stating ‘strongly agree’ or ‘agree’

5 = proportion of respondents stating ‘very important’ or ‘important’

Questions regarding membership of the European Union were not asked in Georgia and

Russia due to non-membership. Croatia joined the EU in 2013, therefore the historical

importance question was not asked.

In general, locations in Spain, Portugal and Germany, have higher levels of interest in

European issues, and respondents in those nations have more positive views of EU

membership. Russia and Slovakia, in general, are less interested in European Issues. There

are contrasting views in the two locations in both the UK and Finland. Coventry (UK) and

Kuopio (Finland) are generally more positive about European issues compared with

Nuneaton (UK) and Lieksa & Nurmes (Finland).

Literature

European integration began in the 1950s with the Schuman Plan and the formation of the

European Coal and Steel Community (ECSC) (Alter and Steinburg, 2007). A common market

for coal, iron-ore and scrap-metal established in 1953 was successful and started the move

for more ambitious goals economic integration including agriculture, and in 1958 the

European Economic Community (EEC) was launched (Alter and Steinburg, 2007). Since its

inception in the mid-1980s, the European Union (EU), together with its member states, has

undergone a major process of transformation (Schmidt and Radaelli, 2004). The EU has

taken momentous steps, with the development of the ‘single market’ in 1992, the European

Monetary Union (EMU) in 1999, and the continued enlargement, expanding to 28 Member

states when Croatia joined in 2013. New policies and practices have led to social and

economic benefits as well as a developing polity. For all the benefits, there are significant

tensions between the national and EU levels (Schmitt, 2009), there are questions about

legitimacy, accountability and democracy within the EU (Schmitt, 2009; Bogdanor, 2007;

Nacarino et al., 2012) and more recently the euro crisis has meant that the ‘European

Project’ has significant challenges (Pew Research Center, 2013). These challenges have

disproportionately affected young people, with rising youth unemployment or cuts in

education budgets (Horvath and Paolini, 2014).

Each member state have their own view and perspective on the purpose and future role of

the EU. There are a number of ‘visions’ of the EU, which have emerged from these views

and perspectives. Schmidt (2009:24) identified four basic non-mutually exclusive discourse;

MYPLACE August 2014


a pragmatic discourse about the EU as a borderless problem-solving entity ensuring

free markets and regional security

a normative discourse about the EU as a bordered values-based community

a principled discourse about the EU as a border-free, rights-based post-national

union

a strategic discourse about the EU as global actor ‘doing international relations

differently’ through multilateralism, humanitarian aid and peace-keeping.

European countries benefit significantly both economically and politically from EU

membership. EU membership provides access to a single market with more than 500 million

people and the EU size enables lucrative trade deals with other partners from around the

world. Free movement of labour has changed patterns of socialisation of young people and

increased the diversity of European countries (Horvath and Paolini, 2014).

The European Union consists of the European Parliament (EP), the Council of Ministers and

the Commission. This complicated structure differs for national polity and citizens have

expressed concerns that the EU is too technocratic and ineffective (Bogdanor, 2007:14),

questioning the legitimacy of the EU, because of pressure put on national politics due to EU

decision making ‘for the people’ is characterised by ‘policy without politics’ (Schmidt,

2009:14). These issues had led to increasing tensions between the national and EU levels,

with “economic decision-making has increasingly moved upwards towards the EU level

while politics and identity have largely remained national, along with the mechanisms of

electoral sanctions” (Schmitt, 2009:). In the current climate, national politicians are less and

less able to address problems and influence policies in the national context due to

internationalisation (Horvath and Paolini, 2013:4). Member states continue to turn to the

EU when seeking solutions to problems that cannot be solved nationally (Micossi, 2008).

The European parliament are elected democratically. Over the years there has been falling

turnout for elections at a time when the EP powers have been expanding (Bogdanor, 2007),

leading to a Democratic Deficit (Norris, 2011, Kohler-Kohl, 2010, Bogdanor, 2007). The

‘democratic deficit’ is when the perceived democratic performance diverges from public

expectations (Norris, 2011) with consequences leading to a crisis of legitimacy. The turnout

for the May 2014 elections was 43.09%, but this varied between 13% (in Slovakia) and

58.2% (in Greece), amongst the MYPLACE countries. “Elections to the European Parliament

are often used by voters to punish the party in power in their particular member state.

Sometimes they are used to cast a protest vote, and so antisystem and extremist parties

tend to perform well in these elections”. (Bogdanor, 2007:8) This was the case in the 2014

European elections saw an increase in the vote for the Eurosceptic parties.

Young people have grown-up in times of European integration, and the free movement of

labour has led to social and economic changes. This has assisted the position that young

MYPLACE August 2014


people having positive views regarding Europe than other generations. There is a stronger

sense of European identify amongst young people (Horvath and Paolini, 2014:6), which is

partly due to the supera-nationalisation of contemporary politics (Horvath and Paolini,

2014:6) and the emergence of multiple citizenships replacing the traditional single national

citizenship (Spiro, 1997). This together with greater social diversity increases the political

differences within the current generation of young Europeans (Sloam, 2013).

15.1 Level of interest in issues to do with Europe (Q1_5)

This question is examined in Chapter 4: Interest in Politics as part of a cumulative index. In

this chapter the question is analysed as an individual question; ‘How interested would you

say you are in issues to do with Europe?’, with four response options; ‘very interested’,

‘quite interested’, ‘not very interested’ and ‘not at all interested’.

MYPLACE August 2014


Q1_5: Level of interest in issues to do with Europe


The graph is ordered by combining the ‘very interested’ and the ‘quite

interested’ responses.

Graph 15.1: Level of interest in issues to do with Europe by location

Narrative


respondents who are very interested, 42% quite interested, 33% not

very interested, and 9% not at all interested in issues to do with

Europe.

Lumiar (Portugal) had the largest proportion of respondents (84%)

who were very interested or quite interested, in issues to do with

Europe. Trnava (Slovakia) had the smallest proportion (36%) who

were interested. Vyborg (Russia) had the largest proportion of

respondents (31%) who stated that they were not at all interested in

issues to do with Europe.

There is clustering of locations by country, when ordered by ‘very

interested’ and ‘quite interested’, both locations in Portugal and

eastern Germany are in the top third. Both locations in Slovakia,

Russia and Croatia are in the bottom third. There are large differences

between the level of interest in issues to do with Europe between

Coventry and Nuneaton in the UK and Kuopio and Lieksa & Nurmes in

Finland.

MYPLACE August 2014


15.2 Level of Interest in the European Union (Q5_6)

This question is examined in Chapter 4: Interest in Politics and as part of an index. In this

chapter the question is analysed as an individual question, using an eleven-point scale

where high values indicate higher degree of interest in the European Union.

MYPLACE August 2014


Q5_6: Level of Interest in the European Union




degree of interest in the European Union.

Graph 15.2: Mean Level of Interest in the European Union by

location

Narrative

The overall mean for level of interest in the European Union is 5.45

(n=16,787 sd=2.75). This varies by location, with the highest mean

score at Telavi (Georgia) with 7.54, and the lowest mean score at two

locations in Russia, Kupchino (3.95) and Vyborg (3.97).


Portugal, Spain and Greece all in the top third, showing higher interest

in the European Union. Russia, Slovakia and Denmark are clustered in

the bottom third.

Most countries had similar mean scores between the two locations;


1.54 difference between two mean scores, Finland (Kuopio and Lieksa

and Nurmes) with 1.27 and Latvia (Agenskalns and Forstate &

Jaunbuve) with 1.20 difference between level of interest in the

European Union between two locations.

MYPLACE August 2014


15.3 Level of trust in the European Commission (Q7_7)

This question is examined in Chapter 5: Attitudes and Trust together with respondents’ trust

in 13 different institutions and organisations. The level of trust in the European Commission

was assessed using 0-10 scales. This approach of examining trust towards European

institutions is often used in several surveys like Eurobarometer or European Social Survey

(Roth et al., 2013).

Table 15.3 Percentage of Missing Values by question

Country Trust in European

Commission

Croatia 7.8

Denmark 13.3

Estonia 12.4

Finland 14.7

Georgia 12.6

Western Germany 11.3

Eastern Germany 5.9

Greece 5.9

Hungary 10.7

Latvia 4.5

Portugal 3

Russia 13.1

Slovakia 7.9

Spain 7.4

UK 14.8

Total 10.40%

MYPLACE August 2014


Q7_7: Level of trust in the European Commission




trust in European commission.

Graph 15.3: Mean Level of trust in the European Commission by

location

Narrative

The overall mean for trust in European Commission is 5.17 (n=15,333,

sd=2.39). This varies by location, with the highest mean score at Telavi

(Georgia) with 6.60, and the lowest mean score at two locations in

Greece Argyropouli (3.45) and New Philadephia 3.66.


Denmark, and eastern Germany all in the top third, showing higher

trust in European commission. Greece, Croatia and Hungary are




0.96 difference between two mean scores, Latvia (Agenskalns and

Forstate & Jaunbuve) with 0.86 difference and Estonia (Narva area

and Tartu) and Finland (Lieksa & Nurmes and Kuopio), both showing

0.80 difference between mean score trust in European commission

between two locations.

MYPLACE August 2014


15.4 Membership of the European Union greatly benefits

this country (Q26_5)

This question is examined in Chapter 9:Citizenship. Respondents were asked if they agree or

disagree to the statement, ‘Membership of the European Union greatly benefits this country

(where applicable)?’ Respondents had five options; ‘strongly agree’, ‘agree’, ‘neither agree

nor disagree’, ‘disagree’ and ‘strongly disagree’ when answering this question.

Respondents from 13 countries answered this question (excluding Russia and Georgia).

Respondents in Croatia were asked this question in the period leading up to becoming EU

members in 2013.

MYPLACE August 2014





responses.

Graph 15.4: Membership of the European Union greatly benefits this

country by location

Narrative


respondents who ‘strongly agree’, 47% ‘agree’, 32%’ neither agree or

disagree’, 11% ‘disagree’ and 2% ‘strongly disagree’ that membership

of the European Union greatly benefits this country.

Bremen (western Germany) had the largest proportion of

respondents (73%) who ‘strongly agree’ or ‘agree’ that membership of

the European Union greatly benefits their country.

Rimavska Sobota (Slovakia) had the smallest proportion (26%), who

‘strongly agree and agree’. New Philadelphia (Greece) had the largest

proportion of respondents (39%) who stated they ’strongly disagree

or disagree’ that membership of the European Union greatly benefits

their country.

There is clustering of locations by country, when ordered by ‘strongly

agree’ and ‘agree’, both locations in eastern Germany, western

Germany, Spain and Portugal are in the top third. Both locations in

Hungary, Greece and Croatia are in the bottom third.

MYPLACE August 2014


13.5 Historical importance of the country having

membership in the European Union (Q51_9)

This question is examined in Chapter 14: History and Memory, together with a number of

other historical events. This question was asked in relation to historical significant events

and was measured on a five-point scale; ‘very important’, ‘important’, ‘neither important

nor unimportant’, ‘not very important’ and ‘not important at all’.

Not all countries are members of the European Union, therefore respondents from Croatia

(who joined in 2013, but did not have an historical context), Georgia and Russia were not

asked this question. Responses are based on 12 countries and 24 locations.

MYPLACE August 2014


Q51_9: Historical importance of the country having membership in the European Union


The graph is ordered by combining the ‘very important’ and the

‘important’ responses.

Graph 15.5: Historical importance of the country having membership

in the European Union by location

Narrative


respondents who said ‘very important’, 45% ‘important’, 11% ‘neither

important nor unimportant’, 3% ‘not very important’ and 1% ‘not

important at all’, that membership in the European Union has been in

the history of each country.


stating that it was ‘very important or important’. Rimavska Sobota

(Slovakia) had the smallest proportion (23%) stating it was

‘important’.

There are clustering of locations by country, with both locations in

Spain and Denmark in the top third. Both locations in Slovakia, UK and

Hungary are in the bottom third.

MYPLACE August 2014



Multi-level models dealing with European issues can be found in the following chapters;

Chapter 5: Attitudes and Trust includes a multi-level model ‘Model B: Trust in European commission’

Chapter 9: Citizenship includes a multi-level model ‘Model В – Membership of the European Union greatly benefits this country’

15.7 Summary

In this chapter, we reported the findings of cross-country analysis of respondents’ attitudes

towards European Issues.

Young people are interested in issues to do with Europe with 58% of young people stating

they are either ‘very interested’ or ‘quite interested’ in European Issues. Young people in

Portugal had the largest proportion of respondents (Lumiar = 84%, Barreiro = 78%) who

were ‘interested’, followed by eastern Germany (Jena and Rostock).

The average level of Interest in the European Union across the respondents was 5.45 (on a

scale from 0 to 10), which is slightly below the mid-point of the scale. The highest mean

score was in Telavi (Georgia) at 7.5 followed by Kutaisi (Georgia) at 7.1. Mediterranean

countries of Portugal, Spain and Greece, together with Germany (both eastern and western)

all have interest levels higher than the mid-point of the scale.

The Average level of trust in the European Commission is slightly higher than in national

parliaments (5.2 on a scale from 0 to 10), although still moderate, with the highest mean

score in Telavi (Georgia, M=6.6) and the lowest mean score in the two locations in Greece

(Argyroupouli, mean=3.5 and New Philadelphia mean=3.7). However, more than half of

MYPLACE countries showed trust in the European Commission at levels above the midpoint

of the scale (Denmark, Finland, Georgia, eastern and western Germany, and Spain and

Portugal). The lowest level of trust in the European Commission is characteristic for youth in

locations within Greece, Croatia, Slovakia and Hungary. Significant differences in young

people’s levels of trust in national parliaments as well as in the European Commission

between locations within countries were more the exception than the rule. In the case of

national parliaments, differences were revealed in three countries (Latvia, Russia and the

MYPLACE August 2014


UK), while in case of the European Commission, differences were revealed in four countries

(United Kingdom, Latvia, Estonia and Finland).

The respondents from the MYPLACE countries who are members of the European Union (all

except Croatia (at the time of survey), Georgia and Russia), 84% stated that it was either

‘very important’ or ‘important’ for the country having membership in the European Union.

References Alter K. J. and Steinburg, D. (2007) “The Theory and Reality of the European Coal and Steel Community” in Meunier, S. and McNamara, K. R . (Eds.) Making History: European Integration and Institutional Change at Fifty. vol. 8 of The State of the European Union, pp. 89–104, Oxford University Press, Oxford; New York.

Bogdanor, V. (2007) Legitimacy, Accountability and Democracy in the European Union. A

Federal Trust report

European Commission (2013) Standard Euroabarometer 80: Public Opinion in the European

Union

Horvath, A.and Paolini, G. (2014). “Political Participation and EU Citizenship: Perceptions

and Behaviors of Young People.”

http://eacea.ec.europa.eu/youth/tools/documents/perception-behaviours.pdf

Kohler-Koch, B. (2010) Civil society and EU democracy: ‘astroturf’ representation? Journal

of European Public Policy 17(1): pp. 100-116

Micossi, S. (2008). “Democracy in the European Union.” CEPS Working Document No. 286,

Nacarino, R. C., De Corte S. and Freudenstein R. (2012) Democracy and Legitimacy in an

Economic Union: Policy Brief. Centre for European Studies

Norris, P. (2011) Democratic Deficit: Critical Citizens Revisited. Cambridge University Press:

New York

Pew Research Center (2013) The New Sick Man of Europe: the European Union


Schmidt, V. A., and Radaelli, C. M. (2004). Policy change and discourse in Europe: Conceptual

and methodological issues. West European Politics, 27(2), 183-210.

Schmidt, V, (2009) Re-Envisioning the European Union: Identity, Democracy, Economy. JCMS

Annual Review. 47: pp. 17–42

Sloam, J. (2013) ‘Voice and Equality’: Young People’s Politics in the European Union, West European Politics, (36)4: pp. 836-858

Spiro, P. J. (1997). Dual nationality and the meaning of citizenship. Emory Law Journal, 46(4):

pp. 141-1485.

http://eacea.ec.europa.eu/youth/tools/documents/perception-behaviours.pdf


MYPLACE August 2014


Chapter 16: Summary

This chapter provides a summary of each of the individual empirical chapters.

16.1 Political Interest: Summary

Following the dominant approach in the literature, which focuses on attitudinal as opposed

to behavioural aspects of political interest, the following definition of political interest was

adopted here: political interest is ‘curiosity, relation, attention and vigilance to political

issues and public affairs‘ (Schmid, 2004).

Political interest has been researched mainly within the framework of political activism. It

has been analysed as a factor, which influences a range of political attitudes, opinions and

forms of political activism. Political interest as a dependent variable, as a phenomenon in its

own right, has received relatively little research attention. There is no thorough

understanding of how it forms and what determines its levels. However, it is clear that the

level of political interest is very stable over the individual lifecourse and that, it forms to a

significant extent in quite early years, being strongly influenced by one’s parents and family.

Later in life, it fluctuates only as a result of very significant social disturbances and changes

such as the fall of Berlin Wall and the collapse of the Eastern Bloc in the late 1980s. Less

significant social and political events do not have sufficient weight to alter its levels

significantly.

MYPLACE data showed that the political interest of young people is divided. Across all

countries, more people reported that they were ‘not interested’ in politics than those who

reported that they were ‘interested’ in politics (58%:42%). The level varied across countries;

young people were interested in politics in some locations (in eastern Germany, Greece and

Spain) while, in other countries, they were not very interested (Croatia, Latvia, Finland and

Estonia). Youth was most interested in politics related to the country they lived in and the

municipality, they lived in. Interest was lowest towards countries neighbouring their

country. Europe and the immediate surroundings were of medium interest.

Employment, housing and environment were the three issues that were of interest for the

highest percentage of young people. The European Union, immigration and LGBT rights, in

contrast, scored lowest on the scale of political interest. Television and the Internet were

the two information channels used by young people to obtain information about politics

and public affairs. Radio and newspapers were far less popular. Levels of different aspects of

political interest vary across sites, but differences usually are not large. When sites are

ordered from highest to lowest then the differences between immediate neighbours are

incremental. In the case of most questions, no single research locations are characterised

MYPLACE August 2014


by exceptionally high or low values of political interest. Neither are there concrete clusters

of sites.

Across all five questions on political interest, certain sites tend to report systematically

lower or higher values. The pattern is far from deterministic and contains exceptions but

one can distinguish between sites where political interest tends to be higher and where it

tends to be lower. To give a very ‘broad brush’ ranking of sites, a cumulative scale summing

responses from the entire five question block was developed and sites were ordered

according to mean value of the index. The following 10 sites tended to have relatively high

political interest:

11. Coventry (the highest average political interest)

12. Odense East

13. Jena

14. Odense Center

15. Vic

16. Bremen

17. Sant Cugat del Vallès

18. Lumiar

19. Rostock

20. Telavi

Among the sites where political interest tends to be low, are:

11. Pescenica Zitnjak

12. Lieksa and Nurmes

13. Nuneaton

14. Narva

15. Vyborg

16. Ozd

17. Trnava

18. Rimavska Sobota

19. Sopron

20. Forstate and Jaunbuve (the lowest average political interest)

This ordering of sites was obtained from summarising responses to all items in the five

questions. This index was not used earlier and thus the ranking of sites might vary from that

presented above. When using separate questions (Q1 to Q5) for ordering the sites, there is

some variation in rankings of the places. However, sites that scored high on one of the

questions, and appeared in the group of sites with high political interest, did not score so

low on any other question such that it also appeared in the group of sites with low political

interest. The opposite was also true. To conclude, despite some variation, the sites tended

MYPLACE August 2014


to group in categories of low and high political interest. Detailed information is presented in

the graphs and tables above.

For multilevel analysis, the cumulative index built of responses to the first question (interest

in five different administrative/political/geographical units) was used. The level of political

interest was to a large extent explained by individual level variables (90%) and the role of

localities is small (10%). The variables, which significantly co-vary with the index of political

interest, are the following:


interest.


Majority/minority group membership: belonging to a minority group means also



political interest.





interest. Interest in politics is positively associated with the level of education. Education is

believed to provide an individual with the skills and knowledge that are necessary for

making sense of political situations and developments and locating oneself in relation to

them. This line of reasoning is confirmed also by the finding that people scoring higher on

the political knowledge index also scored higher on the political interest index. Age too has a

positive association with political interest. This, however, is likely to a large extent to be a

product of education – in this age group, education and age are positively correlated. The

presence of parents’ influence is in line with earlier research, which indicates the influence

of family to be of crucial importance. However, since the level of parents’ political interest

is respondent-assessed, the validity of this measure is hard to ascertain. Reasons for higher

political interest among minority groups need further research. At the present stage, it can

be hypothesised that minorities are somewhat more interested in politics and public affairs.

In this section, we analysed the level of coincidence between the political ideas of young

people and their families and peers. Through this, we can gain an idea of the role of the

environment in the political socialisation process. By measuring the declared closeness of

political views with family and peers, it was shown that, while differences are not very high,

young people tend to feel closer to their peers than their family.

The MYPLACE data set – drawing on 30 different European locations – provides the

opportunity to analyse the role of context in shaping different phenomena. In this case,

however,it seems that context does not have an important role to play in explaining how

MYPLACE August 2014


close or distant the political ideas and values of young people are from their families and

peers.

The analysis of the effect of different variables in defining the closeness of political ideas to

those of family and peers has shown interesting results concerning the role of political

activism. As young people participate more in political actions, they tend to share more

political views with their peers and less with their families. Thus, it seems that political

engagement facilitates the establishment of new political preferences, which are

independent of the family background.

16.2 Attitudes and Trust: Summary

In this chapter, we reported the findings of a cross-country analysis of youth attitudes

towards politicians and politics as well as trust ininstitutions.












scale.

The construction and testing of composite variables based on cross-country merged data

sets revealed that three items related to youth attitudes towards politics and politicians

constitute a one-dimensional scale of acceptable reliability. However, in order to reduce the

share of missing values we used only two items (Q6_2 and Q6_3) for the construction of the

‘Attitude towards politicians and politics’ scale, which also has higher reliability than the

three item scale.

Descriptive analyses revealed that young people in general have very negative attitudes

towards politicians and politics (mean=2.5, on scale from 0 to 8), whereas means scores by

locations range from 1.1 (in New Philadelphia Greece) to 4.6 in Odense Center (Denmark).

Thus, youth in almost all locations on average has negative attitudes towards politics and

politicians (with the exception of locations in Denmark and Finland where the average

MYPLACE August 2014


attitude is neutral). The most negative attitudes towards politicians and politics were

characteristic for youth in Croatia, Slovakia, Portugal, Spain and Greece. Descriptive analyses

did not reveal significant differences in youth attitudes towards politics by locations within

countries.

With respect to analyses related to the construction and testing of composite variables

based on cross-country merged data about youth trust in 13 different institutions and

organisations we calculated two composite scores: overall trust in institutions (9 items) and

trust in national political institutions (3 items), both of which represent good scales.

Besides these composite results we also used one item measures towards specific

institutions: trust in national parliament and trust in European Commission depending on

the topic of further analysis.

Descriptive analyses revealed that young people are, characterised by a low level of trust in

institutions. The overall mean for trust in institutions and organisations is 43.6 (on a scale

from 0 to 90), while the overall mean for trust in national political institutions is 12.5 (on a

scale from 0 to 30). In both cases only in five out of 14 MYPLACE countries do young people

in both locations show trust levels above the midpoint of the scale (in Finland, Georgia,

Denmark, and Germany (eastern and western)). The lowest levels of overall trust towards

institutions as well trust in national political institutions were characteristic for youth in

Greece, Croatia, Spain and Hungary (and Portugal in the case of trust in national political

institutions). The descriptive analyses did not reveal significant differences in young people’s

general level of trust in institutions or their level of trust in national political institutions by

locations within countries.

Regarding the trust in national parliament the results were similar, youth expressed on

average medium levels of trust in national parliaments (mean=4.5 on a scale from 0 to 10),

which varies by location ranging from 2.6 in Pescenica (Croatia) to 6.5 in Odense Center

(Denmark). At the country level, only five of the MYPLACE countries show a level of trust in

the national parliament above the midpoint of the scale (Denmark, Finland, Georgia and

both eastern and western Germany), while the lowest level of trust in the national

parliament is characteristic for youth in both locations within Croatia, Greece and Hungary.

Average levels of trust in the European Commission is slightly higher than in national

parliament (5.2 on a scale from 0 to 10), although still moderate, with the highest mean

score in Telavi ( Georgia, mean=6.6) and the lowest mean score at two locations in Greece

(Argyroupouli, M=3.5 and New Philadelphia, mean=3.7). However, more than half of the

MYPLACE countries showed levels of trust in the European Commission above the midpoint

of the scale (Denmark, Finland, Georgia, eastern and western Germany, Spain and Portugal).

The lowest level of trust in the European Commission is characteristic for youth in locations

within Greece, Croatia, Slovakia and Hungary. Significant differences between locations

within countries in young people’s level of trust in national parliaments, as well as in the

MYPLACE August 2014


European Commission, were more an exception than the rule. In the case of national

parliaments, differences were evident in three countries (Latvia, Russia and UK) while in the

case of the European Commission differences were found in four countries (UK, Latvia,

Estonia and Finland).

In order to explore the multivariate determinants of trust in national parliaments as well

trust in the European Commission, taking into account the nested structure of our data set,

we employed multilevel modelling. Initial results confirmed the justification of multilevel

analyses since it was demonstrated that 17% of the variance of trust in national parliaments,

and 7% of the variance of trust in the European Commission can be explained at the level of

countries. The percentage of variance which could be explained at the level of locations

within countries was much lower: 2% in case of trust in national parliaments; and 2% of the

variance in the case of trust in the European Commission. Such findings are in accordance

with the above and suggest relatively rare significant differences in youth levels of trust in

institutions between the two contrasting locations within countries. Thus, regarding the first

hypothesis according to which youth trust in national parliaments and trust in the European

Commission are differentiated by individual level characteristics as well as location and

country level characteristics, the above findings demonstrate that country level

characteristics are more important determinants of youth trust in national parliaments and

the European Commission than location level characteristics. Additionally they indicate

country level characteristics are more important determinants of trust in national

parliaments than trust in the European Commission. In further analyses we tested several

individual and country level characteristics as potential determinants of trust in national

parliaments as well as trust in the European Commission. Both regression analysis at the

individual level included the main socio-demographic variables (gender, age and social

class) and additional individual level variables, which represent more general dispositions

relevant in the context of trust in institutions such as social trust and life satisfaction, as well

as more politically relevant attitudes such as interest in politics (composite measure) and

satisfaction with democracy. Additionally, analyses of trust towards national parliaments

included a variable on ‘lack of care’ operationalised as agreement with the statement

Politicians are interested in young people like me. These individual level variables explained

29% of variance in trust in national parliaments within countries, and 19% of variance in

trust in the European Commission. This difference probably could be attributed to the fact

that the model with trust in the national parliament contained one more individual level

variable than the model with trust in the European Commission as a dependent variable,

which is, moreover, confirmed as the most important predictor of trust in the national

parliament (the variable concerning politicians’ ‘care’ about ‘young people like me’).

As the individual level determinants of trust in the national parliament analysis revealed, all

socio-demographic variables were significant29; higher trust in national parliament is more

characteristic for females as well as youth with higher and highest social class and younger

29

Due to the large N we are taking into account only variables which are significant on p<0.01

MYPLACE August 2014


youth. In the case of trust in the European Commission the results were similar with regard

to gender and social class; however, age was not confirmed as an important independent

predictor of trust in the European Commission. Among other individual level variables, life

satisfaction, satisfaction with democracy and interest in politics are confirmed as significant

independent predictors of trust in national parliaments as well as trust in the European

Commission. This indicates that young people who are more satisfied with life in general,

and with democracy in their country, are particularly characterised by higher levels of trust

in national parliaments as well as in the European Commission. However, it seems that the

role of interest in politics is different for the case of trust in national parliament and the

European Commission. Namely, in the case of trust in national parliaments higher levels of

trust were more characteristic for those who have greater interest in politics, while, in the

case of trust in the European Commission. those who have greater interest in politics

showed lower levels of trust. These findings might be verified by employing other

operationalisations of interest in politics such as one based not only on self-reported

political interest but also a measure of following politics in the media. Additionally, these

findings could be a result of the presence of an additional individual level variable in the

individual level model of trust in national parliaments (whether politicians are interested in

‘young people like me’).

In the second model we tested the effect of country level variables: welfare state,

corruption perception index and rule of law estimate in the case of trust towards national

parliaments; and welfare state, corruption perception index, government effectiveness and

GDP per capita in analyses of trust in the European Commission. We did not confirm

significant (p<.01) effects for these variables in the case of trust in national parliaments. In

case of the European Commission, only the effect of GDP was significant (p<.01), but this

was negligible.

Finally, in the third model we tested the interaction between one individual level variable

and one country level variable in order to test the hypothesis that the relationship between

individual level determinants of trust in national parliaments and the European Commission

depends on context (specific location or country). In the case of trust in national parliament

our hypothesis was that the effect of an individual’s perception of politicians’ interest in

young people (care about young people like me) as an individual level determinant of trust

in national parliament could depend on a country’s level of corruption (CPI). In the case of

trust in the European Commission our hypothesis was that the effect of left-right political

ideology as an individual level variable could depend on the type of welfare state. However,

we did not confirm these expectations. Thus, it might be concluded, that we did not confirm

the general hypothesis according to which the relationship between individual level

determinants of trust towards national parliament and the European Commission depends

on context (specific location or country). However, it should be stressed that this does not

mean that this general hypothesis is not correct; it only means that it is not confirmed by

those combinations of variables included in the models tested here.

MYPLACE August 2014


16.3 Political Activism: Summary


opportunity and 17% made a conscious decision not to vote. The remaining 11% reported

that they would have voted but were unable to. Locations with voting levels in excess of

80% were Podsljeme, both Danish locations, Kuopio (Finland), both eastern German

locations, Agenskalns (Latvia) and Vic (Spain). By far the lowest level of voting was in

Nuneaton (UK) where only 39% of eligible voters chose to do so, with 43% of eligible voters

deliberately choosing to stay away from the polls.

More than half of respondents (59%) declared that they did not feel close to any political

party.

In this chapter we analysed potential predictor variables of youth political activism. In order

to address this issue we opted to use multilevel regression analysis as MYPLACE data is

organised in a hierarchical structure. Individuals are nested in localities and localities are

nested in countries. One important dimension of political activism is voting behaviour. From

the multilevel analysis performed it was observed that it is in national elections that there is

a higher level of variance between localities, compared to local elections. In other words,

contextual variables have more impact in national elections than in local elections.

The results of variable coefficients allowed us to conclude that predictor variables follow

very similar patterns in the two types of elections. Older voters have higher log odds of

voting. Religion also has an influence on voting behaviour. Those who declared higher levels

of religious identity tend to demonstrate higher levels of participation in local and national

elections.

As seen in previous studies, lower levels of education, belonging to a lower social class and

minority groups decrease the log odds of voting when compared to more privileged groups,

such as those with higher education, higher social class and integrated in the ethnic

majority. Gender is one of the few predictors that appears to behave differently with

respect to national or local elections. Males have higher log odds of voting in national

elections than females, but this difference is not statistically significant regarding local

elections.

According to the literature on political efficacy, political involvement has a positive impact

on turnout rates. Our results support this statement. The increasing of political efficacy and

political involvement levels also increases the log odds of voting. Regarding contextual

variables, we can see that different political systems have different effects on voting

MYPLACE August 2014


behaviour. Liberal welfare state regimes tend to discourage turnout. Bicameral regimes and

proportional representation parliaments increase the log odds of participation in elections.

We have also analysed closeness to political parties among young people. According to the

data, 10% of the variance of youth closeness to political parties is explained by the

differences between localities. A more detailed analysis shows a lower predictive ability of

socio-demographic variables such as employment status, educational attainment or ethnic

belonging; none of these variables has statistically significant capability to predict the

closeness of the participants to a specific political party. However, the same is not true of

social class, it was observed, that belonging to the higher social class increases the log odds

of proximity to a political party.

Political efficacy, political involvement and trust in the political system also have an

influence on party identity; increasing levels of efficacy, political involvement and trust in

the political system result in higher log odds of feeling close to a political party. Regarding

contextual variables only cameral regimes and type of parliament representation influence

the closeness to political parties. Results suggest that living in countries with unicameral

regimes and proportional representation parliaments promote closeness to politics and

decision making and, in this way, to political parties.

The MYPLACE survey allows for extended analysis of non-electoral political actions. This is

crucial when analysing young peoples’ participation since a number of published studies

point to the diversification of the political repertoires of young people (Soler-i-Martí, 2013).

In addition, the MYPLACE data permit the study of the impact of factors at the individual

and the contextual level. This is very interesting given the huge variation in the level of

political activismacross Europe. In order to capture diversity in forms of participation, we

first have to identify different dimensions in the participation repertoires of young

Europeans. It is interesting to note that the first dimension in a factorial analysis identifies a

private and individual form of participation in which the use of the Internet or political

consumerism play important roles. This is the form of participation more prevalent among

young people. It is also the form of participation that is less demanding in time and other

costs such as exposure to the public since it usually occurs in the private sphere. The other

two dimensions of political participation that are identified capture some more classical

dimensions of political participation: traditional, and more institutional, participation (that

we have called public traditional participation); and protest actions (both legal and non-

legal). It is a minority of young people who carry out these two forms of political action.

Nevertheless, especially in the case of protest, there is great variation across countries and

locations.

Consequently, the focus of this chapter has been the explanation of differences across

locations in levels of participation. When we look at the global level of non-electoral

participation, the differences between the 30 locations in the study are huge. As an

MYPLACE August 2014


example, the two locations in Hungary present a level of participation 25 times lower than

some locations in eastern Germany or Spain. Some of these differences are caused by what

is called the compositional effect. That is that the different locations have different

compositions of individuals. However we have also analysed the effect of contextual

variables. We have seen that welfare regimes have an impact on participation and that the

level of membership in organisations improves the participation level of young people in a

location. We also wanted to see the effect of the economic crisis on the participation

patterns of young people. The results for the global indicator of non-electoral participation

show that the impact of the crisis in youth unemployment has a negative effect. However, it

is interesting to highlight that when we analysed protest actions specifically, the opposite

effect is found. Therefore in locations that have experienced a greater impact of the crisis,

young people participate less in global terms but tend to participate more through protest

actions. In these contexts, where the dissatisfaction with institutions is higher, young people

opt for a more confrontational form of participation.

16.4 Civic Engagement: Summary

Consistent with previous studies, our study of young Europeans aged 16-25, has shown that

there is considerable variation across countries and the variation is greater than the

variation between different locations in each of the countries. Similarities with previous

studies are also evident in terms of how the countries are ranked in relation to each other.

Young people from the Nordic countries (Denmark and Finland), clearly have the highest

levels of membership of political associations, while young people from Mediterranean

(Portugal, Spain and Greece) and post-communist countries (Hungary, Slovakia, Georgia and

Latvia) have the lowest levels.

Based on institutional theory, and consistent with previous studies, we suggest that these

differences can be explained best by the ways in which states cooperate and support

political organisations. In the Nordic countries, especially Denmark, many voluntary

associations and organisations receive direct and indirect public subsidies with the purpose

of developing associations and civic engagement. Part of the conditions for receiving

economic funding is that these associations must comply with basic rules regarding

democratic organisation and economy. This also applies to political parties. Also labour

unions are encouraged through the tax deduction that members receive for their

membership.

From an institutional perspective, it is also important how the relationship between the

state and unions has evolved historically. In former communist countries political

associations, which were in opposition to the established regime, weres banned and

citizens’ political participation was only accepted if it supported the communist regime.

MYPLACE August 2014


In relation to individual level variables, a surprising finding was that young women seem to

have a higher rate of membership than their male counterparts. While previous studies

have indicated a trend towards decreasing gender inequality, our research provides

empirical evidence that young women are about to overtake men in terms of their

membership of political organisations.

16.5 Effective Political Action: Summary

This chapter presents analysis conducted on the Q19 items of the MYPLACE survey, which

asks respondents to rate their perceptions of the effectiveness of a series of political

activities in influencing politics in their countries.





The analysis proceeded in three main stages. First, we conducted a factor analysis to reduce

the number of indicators to underlying latent variables. The factor analysis produced a two-

factor solution, with Q19 items tracking on to two separate scales: perceptions of the

effectiveness of legal or nonviolent political activity; and perceptions of the effectiveness of

illegal or violent political activity. We constructed both scales accordingly for use in future

analysis. Second, we conducted a descriptive analysis of all Q19 items and scales by survey

location. Individual items and scales did exhibit clustering by locations, although it is difficult

to identify patterns without statistical tests. In general, voting in elections received the

highest mean response for effectiveness of political action, followed by gaining publicity

through media exposure and being active in a political party. Participating in violent and

illegal protest activities are the least effective.In particular, the two Georgian locations

tended to rate all Q19 items as relatively effective while Russia and the Central European

countries (including Hungary, Croatia and Slovakia) tended to be relatively pessimistic about

the effectiveness of Q19 activities. Third, we conducted a multilevel regression analysis to

explore individual-level and group-level determinants of variation in the effectiveness of

legal or nonviolent activity scale. The results of the model suggest some support for the

hypotheses that the use of the internet for information and political knowledge were

associated with increased perceptions of effectiveness of legal or nonviolent activity.

However, the model provided little support, although did not conclusively falsify, major

hypotheses about youth political activity, including those derived from economic

modernisation theory, and the hypothesis that supranational political and economic change

may lead young people to be more pessimistic about the effectiveness of constructive

political participation. In terms of the goals of the MYPLACE project, this result seems to

MYPLACE August 2014


indicate that accession to supranational political and economic institutions like the EU has

not had any significant effect, either positive or negative, on young people’s attitudes

toward the effectiveness of legal or nonviolent political activity. However, given the

concerns about the model elaborated above, further research is necessary to bring

additional evidence to bear on this hypothesis.

16.6 Citizenship: Summary

Global transformations of contemporary society, the formation of transnational and

multicultural flows and migration processes are changing the traditional understanding of

citizenship as the formal belonging of an individual to the nation state. In this regard, in our

study we focused on various forms of contemporary citizenship, including certain values and

experience of self-determination. Taking into account the whole variety of interpretations

and meanings, we focused on the value assets of national identity, and young people’s

views on global and local state relations in cultural, economic and political dimensions.

Formation of one’s own civic views and position, is inextricably connected with the values of

young people and perception of oneself as a citizen of one’s country. Those who are most

proud of their citizenship status30 are the young people from locations in Georgia, Finland,

Denmark and the UK with levels of pride around 90%. Approximately 60% of the young

people from Vic and Sant Cugat (both locations are in Catalonia) state that they are not

proud of their citizenship. Other locations where levels of pride in citizenship were less than

70% are: the Narva area (Estonia), Bremen (western Germany) Jena and Rostock (both in

eastern Germany), and Forstate & Jaunbuve (Latvia).

In general, in both western and eastern Germany, Denmark and Spain, young people placed

low importance on ascriptive indicators of citizenship. The strongest meaning of ethnic

citizenship was found in Vyborg (Russia). The ethnic role of citizenship is also strongly

supported in all the localities in Georgia and Greece. However, there is a strong difference

between localities in the following countries: Latvia (Agenskalns and Forstate and

Jaunbuve), Finland (Kuopio and Lieksa/Nurmes), Estonia (Narva and Tartu) and Slovenia

30 Our analysis is sensitive to nuances across our research locations where nationality, ethnicity and citizenship




nationality and ethnic status. Overall 80% of respondents can be attributed to the dominant majority30

of the

country, i.e. not belonging to any minority citizen, national or ethnic group within the country though this

varies significantly, in Estonia for example, in the Narva area only 9% are classified as part of the majority

whereas in Tartu the figure is 96%.

MYPLACE August 2014


(Trnava and Rimavska Sobota). In these countries, indicators of support for ethnic

nationalism between the two different localities are significantly different.

In our study there are the two opposing groups of countries: one in which a significant

majority of young respondents fully, or partially, support the policy of protectionism

(Greece, Croatia , Slovakia and Portugal); and a second comprising those who do not agree

with it (eastern and western Germany, Denmark). For example, 77-79% of respondents in

the two localities in Greece support the idea that the state should protect its national

economy by limiting imported goods. In contrast, for example, in eastern Germany, the

proportion was only 12%. Support for the protection of sovereignty ranges from 19% in

Narva (Estonia) to 80% in New Philadelphia (Greece). The idea that the ‘country should

follow its own interests even if this leads to conflicts with other nations’, is supported by the

majority of respondents in locations in Georgia, Greece, Hungary and Portugal.A significant

majority do not agree with this position in Spain, eastern and western Germany.

The highest percentage of those who support restrictions for foreigners, for example, on

their right to purchase land, is in Hungary (from 58% to 64% of the respondents in both

localities). A similar view is supported in Georgia and Slovakia. In general, throughout the

whole dataset, this view is supported by 29% of young people. Thus, for example in eastern

Germany, in Jena, only 1% of people support this view. In eastern and western Germany,

Denmark and Spain, a significant majority did not agree with it.

Finally, an important factor in the understanding of citizenship is the evaluation of

membership of the EU. The overall results of the study demonstrated young people’s

support for such integration (the total percentage of the full or partial agreement was 72%).

The highest rate was found in Bremen (westernern Germany), the lowest - in Ozd (Hungary)

and Rimavska Sobota (Slovakia) - 27%. A significant majority of young respondents positively

evaluate EU membership in eastern Germany, western Germany and Spain. However, in

Greece, Croatia and Hungary young people mostly disagree with it.

If we sum up these trends, we can identify a cluster of countries where most young people

share the idea of the nation state in its classical sense: Greece, Hungary, Slovakia, Portugal,

Georgia. There is also a group of countries where young people prefer other state

formations that are not rooted in the protection of "national interests”.

The availability of foreign films, music and books contributes to the development of national

culture received the greatest support in Coventry (UK). Again, there is a clear difference

between countries where a significant majority of young people see the positive impact of

foreign cultural models on the development of national culture (Portugal, eastern and

western Germany, UK) and countries, where young people do not support this view (Russia

and Georgia).

MYPLACE August 2014


In our further analysis, using Multilevel Modeling, we analysed three models, which, in our

opinion, have the strongest differentiating impact on the approach to citizenship: support

for ‘ethnic nationalism‘; ’the importance of the country’s membership in the EU‘; and ‘the

importance of the protection of national sovereignty’.

The first model was based on the analysis of social factors that lead to young people’s

support for ethnic nationalism. It should first be noted that there is an intersection of

ethnic, civil and cultural indicators of citizenship. Thus, the ability to speak the national

language and its relevance for civic nationalism, is associated with ethnic nationalism.

Contextual characteristics of the position of the country and its well-being increase or

decrease the probability of ethnic nationalism. Of the contextual variables included in our

model, the country's GDP per capita has the highest impact on the importance of ethnic

citizenship: the higher the GDP, the higher the support for ethnic nationalism.

Being a citizen of the country and belonging to the dominant ethnic group increases the

probability of ethnic nationalism. The importance of the status of citizenship for a young

person is in a significant linear relationship with support for the concept of ethnic

nationalism. The lower it is, the lower the importance of ethnic characteristics of citizenship.

There is also a direct relationship between negative attitudes towards minorities and

preference for the ’ethnic concept‘ of citizenship. In contrast, the positive experience of

inter-ethnic interaction and wider inter-ethnic networks reduces the relevance of the ethnic

concept of citizenship.

Young people’s support for a closed image of the nation state, strict control of migration,

protection of the national economy against global economic flows and segregation of

migrants increase the probability of ethnic nationalism. Contrary to our expectations, the

post-socialist status of the country was not significant. However, support for totalitarian or

autocratic state forms reinforces ethnic nationalism.

The socio-demographic characteristics of the respondents (education, gender, age,

employment, social class) have no significant effect on the adoption of the ethnic concept of

citizenship.

The second model was based on the assessment of the importance of EU membership.

Confidence in pan-European or supra-national political institutions has an interesting effect

on the evaluation of membership of the EU. Trust in the European Commission leads to

support for EU membership. However, the same trend is found in relation to the national

political institutions (government, parliament, political parties).

MYPLACE August 2014


Violation of human rights in the nation state only reduces positive attitudes towards

membership of the EU. This may occur due to the general loss of confidence in social

institutions as providers of human rights. A positive attitude towards EU membership not

only increases social capital (employment, social class), but also a subjective view

onmaterial well-being and the overall level of satisfaction with life.

The importance of both civic and ethnic nationalism increases the positive attitude towards

EU membership. Again this confirms the importance of both ’ethnic‘ and ‘civic’ components

in the understanding of citizenship.

At the same time, support for a democratic political system and the lack of anti-migrant

sentiments also leads to the positive assessment of membership of the EU. This may create

the basis for the further construction of transnational identity.

The third model is based on the understanding of the prioritisation of national sovereignty.

Support for the idea of sovereignty is directly related to the xenophobic attitudes of our

respondents. Strong anti-migration sentiment leads to stronger support for the protection

of the ‘national interests‘, such as the need to be guided by public interest despite possible

conflicts with other countries (and this applies both to the political and economic

sovereignty). At the same time, belonging to the dominant ‘group’ and satisfaction with the

status of a citizen of the country increase respondents’ agreement with the dominance of

state interests. The level of trust in the national political institutions also has an interesting

effect; a low level of trust increases the will for political and economic sovereignty

In contrast, the factor of the country's financial well-being is very significant; the more

financially successful the state, the less support for the idea of the national sovereignty.

Moreover, it is macro-economic indicators that are important here - individual poverty or

wealth do not play a key role. Similarly, an increasing level of satisfaction with life in general

reduces the importance of the protection of the sovereign rights of the state.

16.7 Social networks, gender relationships and sexuality:

Summary

The scale expressing the level of trust in society indicated highest levels of trust in Finnish,

western German and Portuguese locations, while Greece, Russia, Slovakia and Georgia had

the lowest mean values for level of trust in society. This scale was also used as a dependent

variable in the multi-level model, which identified respondents from Mediterranean

countries as more likely to develop a high level of trust in society. However, Greece was

located in the bottom third of all participating countries when it comes to mean trust

MYPLACE August 2014


scores. As the survey was conducted in a period when Greece was hit particularly hard by

the Euro crisis and the stability of the government and market were undermined the low

level of trust measured is comprehensible.

When it comes to the frequency of meeting friends in person, the Mediterranean countries

as well as Portugal had the highest rates. However, in all locations the majority reported

meeting their friends at least once a week.

Media technologies were used with higher duration in the UK, Georgia, Russia and Croatia,

while Latvian and eastern German field sites had more respondents using technologies for

shorter periods. However, the survey asked for the duration of technology usage for social

contacts, not for frequency, which might produce different results.

The question battery Q34 was added to a scale indicating the diversity of respondents’

network of friends. Portugal, the UK and Germany (eastern and western) had the highest

mean values. The UK and Germany have a high share of migrants, this might be one reason

for more heterogeneous social networks. Georgian and Hungarian respondents had the

most homogeneous circles of friends.

The most positive attitudes towards women`s role in society and homosexuals were

expressed in the Nordic countries (Denmark, Finland) as well as Spain and Germany.

In Georgia, Greece, Slovakia and Russia, gender based inequality tends to be supported by

the interviewed youths. Rejecting equality for homosexuals was, most often expressed in

Greece, Latvia and Georgia.

Women`s access to abortion is most supported in eastern Germany, Spain and Estonia.

Young people in locations in Russia, Greece and Georgia were the most likely to disapprove

of women’s access to abortion.

The multi-level model testing influences on trust in society revealed that 17% of variance

could be explained at the level of locations. Thus, the context variables were of special

interest. In sum, a country`s welfare state system, the rigidity of the regulation of its market

and juridicial system are decisive aspects when it comes to shaping people`s trust in society.

Additionally, individual economic aspects as well as age play an important role.

16.8 Religion: Summary

Measuring religiosity is a difficult task due to the variety of components that it includes. In

the present survey we used an eleven-point scale where respondents categorised

themselves based on whether they believed they were religious or not. From the

comparative analysis of results, different clusters of countries appeared according to their

MYPLACE August 2014


level of religiosity. Georgia, Greece, Croatia and Russia were part of the cluster which

exhibited high scores on religiosity, whereas eastern Germany, Spain and Denmark

constituted the cluster with low levels of religiosity.

Regarding the frequency of attending religious services, Georgia, Croatia and Greece were

again those countries characterised by high levels of frequency of attending religious

services, whereas Finland, Estonia and Denmark belonged to the cluster exhibiting low

levels of frequency.

Georgia and Greece were the countries in which, overall, higher levels of agreement with

the statement that there is only one true religion were encountered; eastern Germany

demonstrated the lowest levels of agreement. Along the same lines, New Philadelphia in

Greece scored highly in ‘strongly agreeing/agreeing’ with the the statement that good

politicians believe in God. The results of the analysis of responses to that statement showed

that there was significant differences both within and between countries. Eighty three per-

cent of the respondents from Sant Cugat, Spain, ‘strongly agreed or agreed’ that religious

leaders should have no influence on government decisions, whereas 34% of the

respondents in Nuneaton, UK ‘strongly agreed or agreed’ to the aforementioned statement.

Again, three clear clusters of locations were identified, with locations in Spain, Greece and

western Germany agreeing that religious leaders should have no influence on government

decisions. Locations in the UK were in the top third disagreeing with the statement.

Fifty three per-cent of the respondents in Sant Cugat in Spain, ‘strongly agreed or disagreed’

with the statement that none of the major religions have any truths to offer while only 4% of

the respondents in Telavi in Georgia ‘strongly agreed or agreed’ with this statement. Spain,

Finland and Latvia constituted the cluster of countries that agreed more with that

statement, whereas Georgia and UK presented high scores of disagreement.

Regarding the most significant determinants of religiosity, the analysis verified our original

hypotheses. More specifically, gender turned out to be a variable that affects levels of

religiosity, as women appeared more religious than men. Numerous surveys have

repeatedly argued that women are characterised by high levels of religiosity and this

explains the fact that in every piece of quantitative research the gender variable is included

(Stark, 2002:496).

As expected and stated in our original hypotheses, Muslim respondents appear more

religious. Orthodox Christians follow, along with Catholic and Jewish respondents.

Protestant respondents expressed religiousness to a somewhat lesser extent than others

declaring belonging to a faith group, although their religiosity was still significant.

Finally, being a part of a minority group increases one’s level of religiosity. This mainly

applies to immigrants who are equipped with different social and cultural capital compared

MYPLACE August 2014


to the native population and thus try to preserve it and to communicate it to the next

generations in order to maintain their collective memory and identity. Religion in this

process of assimilation plays a crucial role and is vital in the process of crafting an identity.

Religion is not a trait or a characteristic that is strongly associated with a country or a certain

place, and thus helps the people in a country to overcome isolation and marginalisation.

Furthermore being active in religious activities and organisations enables the process of

socialisation and provides people with opportunities for civic engagement and social action.

Finally, when religion is a prominent component of a person’s identity and much of

everyday life is organised around these needs, then it is possible that other differences that

mark an individual’s marginalised place in society, are less pronounced.

16.9 Attitudes towards Minority Groups: Summary

The MYPLACE survey shows that young people’s attitudes towards minorities differ from

country to country, and from locality to locality. Young people in Western European

locations, especially eastern Germany, are least likely to have negative views towards

different minority groups and to prefer limiting access by immigrants to a range of

resources. Young people in post-socialist locations, along with Greek locations, have more

negative attitudes and are more likely to be exclusionist towards immigrants. Nevertheless,

in some countries there are large differences between localities, for example, in Latvia and

Finland.

The most negative attitudes are towards the Roma, particularly in in Hungary, Slovakia,

Greece and Russia where there is a considerable Roma minority, as well as Lieksa and

Nurmes (Finland). Attitudes towards Jews are more positive than negative. The most

negative young people can be found in Ozd (Hungary), Trnava (Slovakia), Aegyroupouli

(Greece), and Vyborg (Russia). Attitudes towards Muslims are similarly positive; however

there are localities where negative views towards them are quite widespread: Vyborg

(Russia), Trnava (Slovakia) and Sopron (Hungary).

Young people in Greek locations are much more likely than others to have xenophobic

attitudes; this should be seen in the context of Greece having taken a large proportion

those fleeing unrest from the ‘Arab Spring’. Limiting immigration would also be supported

by young people in Vyborg (Russia), Bareiro (Portugal), Trnava (Slovakia), and Nuneaton

(UK), whereas German locations are clustered at the bottom of the xenophobia scale.

Hungary, Russia and Georgia are at the top of the welfare chauvinism and exclusion scale;

here young people favour limiting access by immigrants to various resources. Interestingly,

Nuneaton youth in the UK tend to be much more xenophobic, welfare chauvinist and

exclusionist than youth in Coventry.

MYPLACE August 2014


One of the strongest predictors of negative attitudes towards minorities, xenophobia,

welfare chauvinism and exclusionism is ethnic nationalism. Like other studies, we also find

that all these negative attitudes are more widespread among men, and among those who

are less educated, whose families are struggling financially and whose parents do not belong

to the highest social class. This can be linked to a higher level of insecurity and perceived

competition. Moreover, perhaps linked to that, we find that young people who are

employed are more likely to have such attitudes rather than those who are not employed.

Age, on the other hand, does not help to explain negative attitudes towards minorities;

contrary to psychological theories, we find that young people are not more likely to have

such negative views as those who are older.





towards minorities. This relationship is intriguing, and there could be several explanations

for it, such as cultural differences or perceived inter-group conflict or competition.



to hold negative attitudes towards minorities. A frequent use of different media channels is,

however, strongly linked to negative attitudes towards minorities, xenophobia, welfare

chauvinism and exclusionism, which shows that the media are at least partly responsible for

the recent spread of populism and far right attitudes in Europe.

The data also confirms the ‘contact hypothesis’; personal contact with minorities decreases

negative attitudes towards minorities and immigrants. Similarly, overall distrust in people

and the government facilitates all kinds of negative attitudes towards minorities and

immigrants.


are poor also more xenophobic and likvely to have more negative attitudes towards

minorities, but the overall situation in the area matters too; in poorer areas such views are

more widespread. At the same time, the effect of GDP growth during the crisis is

insignificant. It shows that resource stress is mainly experienced based on the economic

conditions young people see around them, not general macro-economic developments.

Xenophobia and welfare chauvinism is more widespread where immigration rates are

higher, lending further support for resource stress and competition thesis. Finally, liberal

regimes are more likely to produce such attitudes. Thus the dismantling of the European



MYPLACE August 2014


16.10 Democracy: Summary

In this chapter, we reported the findings of cross-country analyses of the MYPLACE survey,

in which young people’s understandings of democracy are analysed.

A descriptive analysis of Satisfaction with democracy shows that the overall mean across all

30 locations is exactly in the middle of the scale (5.0), but satisfaction varies from country to

country and between locations. There were clear clusters of locations; with locations in

Denmark, Finland, eastern and western Germany all having higher levels of satisfaction with

democracy. Mediterranean countries including Greece, Spain and Portugal have lower levels

of satisfaction with democracy.

An ordinal by ordinal Gamma correlation of Attitudes towards various forms of political

systems (Q43) suggested that two separate scales were formed, one which focused on

Democratic Political System and another which focused on Autocratic principles. Results

showed that young people in locations in Denmark and eastern Germany had more positive

views towards democratic political systems. Respondents in locations in Russia, Latvia and

Slovakia were less positive towards democratic political systems. Both locations in Georgia,

Croatia, the UK and Portugal have more positive opinions towards a non-democratic form of

political system (with autocratic principles ). Both locations in Germany (western and

eastern), Finland and Spain were more negative towards autocratic principles.

A factor analysis showed that all eight items for question 44: Justification for violence are

loaded under a single factor with excellent reliability (α = 0.928). The overall mean for

justification for violence was generally low (9.7 out of a possible 32). Both locations in

Russia, Croatia, and the UK were in the top third, stating that violence is ‘more justified’,

with locations in Latvia, Denmark, Finland and Estonia stating that violence is ‘less justified’.

Items in question 45: Attitude towards social justice and the economy were analysed

individually, because when combined the reliability of these factors was low, with Cronbach

alpha values less than 0.4. This suggests that there are contradictions between respondents’

ideological views in terms of socialist principles (items 2 and 3), and free-market principles

(items 1 and 4). Both locations in Georgia, Greece and eastern Germany stated that they

disagreed that the unemployed should have to take any job available or lose their

unemployment benefits. Both locations in Portugal, Denmark and Hungary are in the top

third and stated that they should. Most young people in the two Georgian locations stated

that Private ownership of business and industry should be increased. Both locations in

eastern Germany and Greece are clustered in the top third disagreeing that Private

ownership of business and industry should be increased. There are clusters of locations,

with locations in Greece, Spain, Hungary, Russia and Slovakia all in the top third, agreeing

MYPLACE August 2014


that Competition is harmful, it brings out the worst in people. When ordered by ‘disagree’

and ‘strongly disagree’, locations in Denmark and Germany (western and eastern) are

clustered in the top third disagreeing that competition is harmful. There are clear clusters of

locations, with young people’s views on Incomes should be made more equal, with

locations in Greece, and Germany (eastern and western) all in the top third, agreeing that

incomes should be made more equal. When ordered by ‘disagree’ and ‘strongly disagree’,

locations in Denmark, Georgia and Estonia are clustered in the top third disagreeing that

Incomes should be made more equal.

A factor analysis of these items showed that all three items are loaded under a single factor

with excellent reliability (α = 0.866) for Freedom of speech. Both locations in Portugal,

Denmark, Greece and Estonia all in the top third, stating that freedom of speech should

always be allowed. Germany (western and eastern), Latvia and Hungary are clustered in the

bottom third.

There is a large variation of views regarding respect for human rights. Both locations in

Finland, Denmark and Germany (eastern and western) are all ranked in the top third, stating

that there is respect for human rights. Both locations in Georgia, Greece and Slovakia are in

the bottom third when stating there is no respect for human rights. Both locations in Greece

had the highest proportion of young people stating there is ‘no respect at all’ (22% and

24%).

An index ranging from zero to five, where a higher score in the index indicates greater level

of discrimination/sense of insecurity in society, were constructed for the five items in the

question on Self-perceived discrimination (Q48). Overall levels of discrimination are low;

with a mean for sense of insecurity in society of 0.4. When ranked, locations in Croatia, the

UK, Russia, eastern Germany, and Slovakia are all in the top third, with higher levels of sense

of insecurity/ discrimination. Portugal, Hungary and Georgia are clustered in the bottom

third with lower levels of sense of insecurity/discrimination.

Opinions on the death penalty (Q49) vary from country to country, with locations in

Georgia, Denmark and Greece all in the top third, stating that the death penalty can ‘never

be justified’. However, Locations in Hungary, Estonia, Latvia, and Slovakia are clustered in

the bottom third with more respondents stating that the death penalty ‘can be justified’.

In order to explore multivariate determinants of Satisfaction with democracy, taking into

account the nested structure in our data set, locations within countries, we employed

multilevel modelling (MLM). The initial analysis of variance components in the ‘empty’

model revealed that 78% of the variance in young people’s ‘satisfaction’ with democracy

can be explained by differences between individuals and that 23% of the overall variance, in

the ‘satisfaction’ with democracy, lies at the level of localities. The substantial variation at

local level justifies the reason for using multilevel analysis in our data. The model confirmed

MYPLACE August 2014


that young people reporting greater political knowledge, satisfaction with life, trust in

parliament, trust in politicians, general trust in people, and positive evaluation of the

country’s human rights situation have a statistically significant higher level of ‘satisfaction’

with democracy. Young males, those with greater involvement with political activities, older

young people, those having parents from upper social class, and those ‘not religious’

appeared to have a statistically lower level of satisfaction with democracy. At a contextual

level, the youth unemployment rate of a country did not appear to have a statistically

significant relationship to young people’s ‘satisfaction’ with democracy, but welfare state

type did have a significant association with ‘satisfaction’ with democracy. Compared to

those from post-socialist countries, young people from Nordic and Conservative type

countries appeared to have significantly greater ‘satisfaction’ with democracy.

16.11 History and Memory: Summary


and 56% a little interested . Respondents consider World War II and membership in the

European Union the most important from their countries’ perspectives among the listed

events and periods, and NY terrorist attacks September 9th 2011 and the Holocaust the

least important.










Our hypotheses concerning the independent variables detailed in the introductory section

are verified by the outcomes of regression models.

The locations, socio-demographic and political characteristics variable’s impact politics-

related activities the most (22%) and the importance for commemorating the past the least

(17%).

Among locations, the Lieksa/Nurmes Finnish location did not have significant effect in any

of the three models. The impact of Tartu (EST) and Bremerhaven (GER-W) was not

significant in the first two models, and the same applies to the Narva area (EST), Bremen

(GER-W), Podsljeme (CRO) and Jena (GER-E) with regard to Model 2 and Model 3, while

MYPLACE August 2014


Agensalns (LAV) was not significant in Model 1 and Model 3. We can state that the

perception of history and the past is, influenced the least in the Finnish, Estonian and

western German locations (non significant effect).

Three locations had significant roles: Rostock (GER-E) had significant positive beta values in

Model 1 and Model 2 and the same applies to Nuneaton (UK) and Kupchino (RUS) with

regard to Model 2 and Model 3. We can conclude that we are more likely to find a higher

level of importance attached to historical events and commemorating the past if

respondents live in any of these three locations.

According to our results, two socio-demographic variables determine respondents’ relation

to the historical past. Parental social class and belonging to the majority society significantly

increase the importance of historical events and commemorating the past. The sex variable

had a significant effect in Model 1 and Model 2 but women tend to attache importance to

historical events while men favour history-related activities.

The perception of the importance of history, the past and historical

memory/commemoration is clearly influenced by political relations; the level of political

knowledge, interest in politics and the support of mainly soft political actions as a means for

influencing politics had significant effects in all three models. A major finding is that past-

related activities participated in characterise those the most who are dissatisfied with

democracy and are right-wing. In contrast being left-wing increases the likelihood of

attaching a higher level of importance to history and commemoration.

16.12 European Issues: Summary

In this chapter, we reported the findings of cross-country analysis of respondents’ attitudes

towards European Issues.

Young people are interested in issues to do with Europe with 58% of young people stating

they are either ‘very interested’ or ‘quite interested’ in European Issues. Young people in

Portugal had the largest proportion of respondents (Lumiar = 84%, Barreiro = 78%) who

were ‘interested’, followed by eastern Germany (Jena and Rostock).

The average level of Interest in the European Union across the respondents was 5.45 (on a

scale from 0 to 10), which is slightly below the mid-point of the scale. The highest mean

score was in Telavi (Georgia) at 7.5 followed by Kutaisi (Georgia) at 7.1. Mediterranean

countries of Portugal, Spain and Greece, together with Germany (both eastern and western)

all have interest levels higher than the mid-point of the scale.

MYPLACE August 2014


The Average level of trust in the European Commission is slightly higher than in national

parliaments (5.2 on a scale from 0 to 10), although still moderate, with the highest mean

score in Telavi (Georgia, M=6.6) and the lowest mean score in the two locations in Greece

(Argyroupouli, mean=3.5 and New Philadelphia mean=3.7). However, more than half of

MYPLACE countries showed trust in the European Commission at levels above the midpoint

of the scale (Denmark, Finland, Georgia, eastern and western Germany, and Spain and

Portugal). The lowest level of trust in the European Commission is characteristic for youth in

locations within Greece, Croatia, Slovakia and Hungary. Significant differences in young

people’s levels of trust in national parliaments as well as in the European Commission

between locations within countries were more the exception than the rule. In the case of

national parliaments, differences were revealed in three countries (Latvia, Russia and the

UK), while in case of the European Commission, differences were revealed in four countries

(United Kingdom, Latvia, Estonia and Finland).

The respondents from the MYPLACE countries who are members of the European Union (all

except Croatia (at the time of survey), Georgia and Russia), 84% stated that it was either

‘very important’ or ‘important’ for the country having membership in the European Union.

MYPLACE August 2014


Appendix A: Guidance Document (preparing combined data

for the thematic report)

MYPLACE

MYPLACE WP4 D4.6: Europe wide thematic reports

Guidance Doc (preparing combined data for the thematic report)

1. Background

This document is related to the Deliverable 4.6 (the last deliverable under WP4) for which

we need to produce a Europe wide thematic report. Each team has been allocated a theme

(including the variables associated with that theme) in our meeting at Slovakia. Each team

needs to work on the assigned theme by using the combined survey dataset. Since this

deliverable requires us to use all data across the participating countries, we need to pay

some special attention in terms of data preparation and analysis.

The purpose of this brief document is to describe the basic procedures we need to follow to

prepare the data before carrying out statistical analysis on each assigned theme. Here we

would like to emphasis the following two aspects:

This doc does not suggest the kind of statistical analysis (univariate, bi-variate, and

multivariate) each partner undertakes for their particular theme.

We anticipate that where teams have the statistical expertise that they will

undertake multivariate modelling, particularly in such a way as to use the case

study aspect of the research design. They may not need help from this guidance

doc as they routinely follow these procedures (and others) as part of standard

practice for cross-country/location wise data analysis. This doc is prepared for less

advanced users. For that reason, we have kept the procedures as simple as possible

and in some cases we have kept it to the minimum.

2. Three-step procedure for preparing data

In this section, we describe a three-step procedure that we can follow as part of preparing

data for cross-European analysis on each theme.

MYPLACE August 2014


Step-1

Although the combined dataset all partners receive will contain harmonised variables, it will

be useful to do a quick check on the following aspects in the combined dataset. In this

regard, partners do not need to examine all variables in the dataset. Instead, they need to

focus only on those variables which they will use in their analysis for the assigned theme.

Coding: Codes for a variable are consistent across all countries in the combined

dataset

Missing values: Missing values are consistent and defined accordingly across all

countries in the combined dataset

Code range: There is no out of range code (as per the code plan) for a variable across

all countries in the combined dataset

What to do if problem exist?

It is to be emphasised that we all have checked those aspects before as part of cleaning data

(Deliverable 4.5) for individual country. We have followed a common strategy for that

purpose and therefore, we normally do not expect these issues to arise in the combined

dataset. However, if we notice any inconsistency in the combined dataset, we need to

identify the country, contact them, and resolve these issues with them before carrying out

any analysis. It might be useful to let CRRC know these issues and the solutions you are

taking so that Tina can amend them (if required) in the ‘master copy’ of the combined

dataset.

Step-2

We know that most of the questions in the survey are commonly asked to all countries.

However, there are some situations where national variance was allowed (for example,

items 7 and 8 from Q40). As part of preparing combined data for cross-national analysis, we

need to identify (if any) those variables (under a particular theme) which are applicable to a

specific country. As per the decision taken in the meeting in Slovakia, the partners need to

drop (if any) these variables from the analysis of data.

Step-3

Partners are encouraged to check the potential for developing scales/indices from the

variables they will be analysing under the assigned theme. This will help reduce the number

of variables they require in the final analysis. It is noted here that in the analysis of data for

the country specific report (Deliverable 4.5), all partners have examined this aspect and

developed scales/indices from their own country data. Partners need to do the same here.

MYPLACE August 2014


The only difference is that now they need to develop scales/indices on the combined

dataset. For this purpose, partners are suggested to consult the guidance doc titled

['Construction of scales/indices' (Background doc—UK Data Analysis)] which we circulated

to them as part of the guidance for Deliverable 4.5.

For a quick reference, we highlight below the main points for constructing Likert scales (see

the guidance doc for further detail):

Checking uniformity of the code for an item under a potential scale across all

countries. This should not be a problem as we all have maintain a common coding

scheme. If problem exists, coding for scale items needs to be done uniformly on

combined dataset.

Making sure all positively and negatively worked items under a potential scale are

coded correctly (see previous guidance doc for further info) and uniformly across all

countries. If problem exists, coding for scale items needs to be done uniformly on

combined dataset.

Running factor analysis on the combined dataset to examine factor structure and

checking internal consistency of items under a factor by Cronbach Alpha. In this

regard, we may have three different situations:

a) Items under a factor can be described meaningfully (by a name/concept) and

items have good internal consistency (measured by Cronbach Alpha with a

minimum value of 0.70). In this situation, we create a summated scale (by

adding scores for all items under a factor). We use this scale in the

subsequent cross-European analysis.

b) Items loaded under a factor can be described by a meaningful concept but

internal consistency of the items are not satisfactory (Cronbach Alpha value

less than 0.70). In this situation, we can think about three possible solutions:

drop those items from the subsequent analysis (if too many variables for

the team to consider under a theme, this option might work better),

use individual item in the subsequent analysis (may be a good strategy if

it relates to an important variable in the study and/or teams have only a

few variables under their assigned theme), or

calculate factor scores (can be obtained by choosing an option available

whilst running factor analysis) and use these scores in the subsequent

analysis. It is to be noted that although factor scores are very useful for

MYPLACE August 2014


running multivariate analysis, they may not be as informative as the

summated scale if they are to be used for calculating measures of central

tendency (e.g. mean) for comparing countries.

c) Items loaded under a factor (or just one item under a factor) cannot be

described meaningfully (irrespective of the level of internal consistency).

Two possible solutions: (a) drop those items from the subsequent analysis (if

too many variables under a theme, this option might work better), or (b) use

individual item in the subsequent analysis (may be good strategy if it relates

to an important variable in the study and/or only a few variables available to

explore under a particular theme).

In addition, as part of analysing data for country specific report (Deliverable 4.5), partners

have constructed a number of indices. For the thematic Europe wide analysis, partners are

also suggested to consult the same guidance doc and construct similar type of indices (from

variables under their assigned theme where possible) from the combined dataset.

3. Main points and future analysis

The steps described above will help preparing data, which will be required for conducting

statistical analysis of variables (under a theme) at a later stage in the combined dataset.

However, we need to acknowledge that the development of a scale (and/or examining its

psychometric properties)/indices is not as simple and straightforward as we have described

above. More sophisticated analysis including the testing of validity may be required for

publishing results in referred journals. More specifically, the analysis suggested here for

examining factor structure and reliability of scales could be extended to individual country

(including two locations in a country) level data. This would allow us to examine how and to

what extent factor structure and reliability of a scale obtained from the combined data

varies from the country or location specific dataset. Therefore, anybody interested to

develop more robust scales/indices by further extending the analysis are most welcome.

However, taking into account a range of factors (audience of this report, time, resource

etc.); we set a minimum threshold described in this doc, which we all aim to achieve for

maintaining quality of analysis in this report.

Political Activism -in Pollock G. and Ellison M. (2014) 'Measuring Participation', European...

Documents

Transcript of Political Activism -in Pollock G. and Ellison M. (2014) 'Measuring Participation', European...