Does reading make you surf more? - Patterns and relationships between commuter's daily reading...

22
J M R Quartley 022642 ED3000 1 Copy 2 Does reading make you surf more? -Patterns and relationships between commuter’s daily reading habits and Internet usage.

Transcript of Does reading make you surf more? - Patterns and relationships between commuter's daily reading...

J M R Quartley 022642 ED3000

1

Copy 2

Does reading make you surf more? -Patterns and relationships between commuter’s daily reading habits and Internet usage.

J M R Quartley 022642 ED3000

2

Abstract

This study compares data for two groups of subjects observed as either

readers or non-readers on morning commuter Underground trains. The study

investigates the cognitive effects of differential exposure to print by assessing

the groups reading practices at home, work, and use of the Internet. The

results showed that subjects observed as readers significantly out-perform

non-readers across a range of literacy and Internet measures. It also showed

that background variables of age, gender and social class are an indicator of

reading and Internet competence. An individual reader’s exposure to print

predicts other literacy practices and characteristics, and supports the

hypothesis that engaging in reading activities leads to the informal

development of new literacy skills.

J M R Quartley 022642 ED3000

3

Introduction

Much has been written around the subject of gaps in access, for individuals

and communities, to information technology –widely referred to as the ‘digital

divide’. Explanations, accounting for this, have mostly been based around four

groups of factors: economic, cultural, psychological and institutional (Neice

1998; Kvasny 1998; Schön, Sanyal and Mitchell 1998). These factors go

someway in explaining digital participation in terms of social

exclusion/inclusion or barriers and gateways to access. However, they do not

describe digital participation as a logical progression in literacy development.

In recent years there has been a shift within the area of literacy studies, where

attitudes have moved away from ‘considering literacy as an autonomous set

of acquired skills to investigating literacy as a social practice’1. Street (2001)

describes literacy as a socially situated practice, where our uses of reading

and writing define our identity and mediate our connections with our

surrounding world, leading to ‘quite new ways of understanding and defining

what counts as literacy.’2 The UK Government’s Essential Skills Strategy

(2002)3 defines literacy as ‘the skills and abilities needed by all to ensure that

they participate in society and working life’. Current demands in the domains

of work and education stress the increasing need for individuals to attain

computer literacy. The rise in home computer ownership, burgeoning

magazine titles offering help advice and consequent Internet access,

suggests that individuals are engaging in informal learning processes to

extend their literacy abilities. Barton and Hamilton (1998) describe these as

‘vernacular strategies’ used to learn new literacies.

Researchers have been interested in the effects of individual reading practice

and the development of reading abilities, which has led to the recognition that

reading practice is rooted in social participation and that social role has an

1 Anne Kernaghan Literacy as a social practice: a case study. Learning and skills development agency. Northern Ireland. 2 Street (2001: 18), see Bibliography/References. 3 As described in: Anne Kernaghan Literacy as a social practice: a case study. Learning and skills development agency. Northern Ireland.

J M R Quartley 022642 ED3000

4

influence on adults reading interests and habits (Gray & Rogers, 1956). The

work of cognitive psychologists, interested in the socio-cultural perspectives of

learning and cognition, proposes that all individual cognitive abilities

originate as internalizations of social interactions in surrounding social

environments (Leontiev 1975, and Vygotsky 1978). This has led to the belief

that literacy activities lead to intellectual benefits, as Smith (1994) states:

Those who have opportunities to read (i.e., through availability of printed

materials, appropriate literacy models and instruction) can develop and refine

their skills; well-developed skills, in turn, enable the reader to accomplish

reading tasks of increasing complexity.

Despite general belief in the positive relationship between reading activities

and the informal development of cognitive skills, only limited research has

been conducted in to this area, and with mixed results (Hayes, 1988;

Stanovich & Cunningham). West, Stanovich & Mitchell (1993) conducted

research into an exposure hypothesis and found evidence to support their

view that widespread ‘exposure to print’ (that being the quantity of an

individual’s daily reading activities) leads to cognitive gains. Further research,

based on this exposure to print paradigm, by Smith, Elliot & Hutchinson

(1994) was ‘unable to distinguish high exposure and low exposure adult

readers on cognitive outcomes presumed to be related to how extensively

people read’.

This study follows the same exposure hypothesis to investigate the

relationship between an individual’s daily reading practice and use of the

Internet. For the purpose of this study Internet use is chosen as a proxy to

describe the intermediation processes associated with the wider class of

digital technologies (Neice 1998) and also because it is predominantly a text-

based media which shares much with reading in terms of the richness of its

information content. Although assessing the informal development of new

literacy skills is beyond the scope of this paper, it is possible to reveal

correspondences between the literacy practices of reading (newspapers,

J M R Quartley 022642 ED3000

5

books, magazines, personal documents...etc) and use of the Internet for

individuals, across age, gender and socio-economic groups.

Given the mixed results from previous ‘exposure’ research and lack of studies

that directly compare reading and Internet usage in this way, it was not

entirely clear what patterns would emerge from the data. However, the

research by West, Stanovich and Mitchell (1993) suggests ‘a role for reading

experience in a comprehensive theory of cognitive growth’, while Smith (1994)

states that such exposure hypothesis studies ‘suggest that exposure to texts

provides practice at literate activities…and promotes skill in those activities’.

Conclusions of research relating to the effects of information technology and

television on reading are similarly mixed: the Programme for International

Student Assessment’s (PISA) results for Canadian students showed a

positive association between reading scores and the frequency of computer

use and home computer access; and Robinson et al (2000) found that despite

Internet usage time tripling from 1995-1999, the amount of time users devote

to other media (Newspaper reading and Television watching) has not been

significantly impacted on. However, Neuman & Celano, (2004) observed

student activities, on computers, and found that instead of using computers as

a tool to strengthen reading skills, they used them in non-academic ways.

Gadberry (1980) found that reducing the amount of television children

watched by half improved Performance IQ and increased reading time.

Despite these latter negative research results, Hayes & Ahrens (1988) note

that the Internet does not exhibit the lack of depth with which television and

radio are associated.

The design of the questionnaire was influenced by the methodological

approaches of two pieces of research:

o The US National Adult Literacy Survey (NALS)4 as described by Smith,

1994)5. A document-type classification from this study was used to

4 Kirsch, Jungeblut, Jenkins & Kolstad, 1993, referenced in Smith M C (1994), 5 Smith M C (1994) What Do Adults Read and Wny Does It Matter? Northern Illinois University. Paper presented at the annual meeting of the Mid-Western Educational Research Association, Chicago, October 1994. http://www.cedu.niu.edu/~smith/Unpubs/mwera94_2.pdf

J M R Quartley 022642 ED3000

6

gather data for individual’s self-reported levels of reading activities in

two social contexts of home and work.

o Stanovich & West’s (1989) detection logic Recognition Tests, as a

checking device against the tendency for individuals to give socially

desirable responses in self-reported data.

The study was conducted in London Underground morning commuter trains,

for the same reasons West, Stanovich & Mitchell (1993) selected passenger-

waiting lounge at an airport as a setting because:

Reading occurs via the free choice of the subject. Experimenters do not

intrude upon the process…[and]…the setting…directly addresses

concerns about ecological validity.

For the purpose of this study, subjects observed in a setting where they are

free to make a choice to read or not, are classified as readers or non-readers

and this distinction is used as a means of comparison, in addition to the

variables of social class and gender, when considering patterns between

reading activities and Internet use.

Methodology Procedure

The study took place on London Underground trains, during commuters’

morning travel to work and was conducted by the author of this research

paper. Subjects were selected as solitary travellers not engaged in

conversation. Before being approached, potential informants were observed

and classified as either readers or non-readers. A reader was defined as

someone engaged in a literacy practice involving reading, a non-reader

defined as not engaging in any reading activity. For readers, a note was taken

of the reading activity that they were engaged in, from the following

classifications: newspaper, fiction, non-fiction, personal papers, religious,

J M R Quartley 022642 ED3000

7

maps, adverts, crosswords, diary, notes, official forms, personal digital

assistants, mobile phones. The gender of the observed reader/non-reader

was also recorded at this stage. Once this initial data had been recorded, the

observed individuals were approached and invited to complete a

questionnaire with the purpose of eliciting certain information about their

reading and Internet practices (Moser and Kalton 1971:271), having first been

described the purpose of the study.

Study environment As noted, the study environment addresses concerns about ecological

validity, but also offers the opportunity to provide a representative sample of

socio-economic classes (Sankoff 1980c:52). Data is collected on two morning

commuter journeys6 in order to incorporate social classifications, using the

National Readership Survey (NRS) classification of social grades7. It is

assumed that commuters travelling into the financial centre of London will

provide data for: A –upper middle class, high managerial, administrative or

professional; B –middle class, intermediate managerial, administrative or

professional; and C1 –lower middle class, supervisory or clerical and junior

managerial, administrative or professional. Correspondingly, commuters

travelling out and away from the financial centre will provide data for the social

grades: C2 –skilled working class, skilled manual workers; D –working class,

semi and unskilled workers; and E –casual or lowest grade workers. Since

this is a small study the target sample size is six respondents for each social

grade (Milroy 1987).

Preliminary investigation noted that there would be serious structural

limitations to gathering data. To avoid crowded train conditions the data would

have to be collected before 8am, since after this time peak periods of travel

6 The first journey for observation and questionnaire conducted on a Northern Line train, between Clapham South and Bank stations. Second journey on the District Line between Bank and Barking 7 A full list of National Readership Survey definitions of social grades can be found in Appendix 2.

J M R Quartley 022642 ED3000

8

would have prevented the researcher from free movement within the carriage.

At the best of times, the presence of a researcher attempting to conduct a

questionnaire, in the characteristic silence that commuters exercise on

London Underground trains, pointed to the possibility of a high rate of refusals

to participate from potential informants. Early pilot observations of literacy

activities on board trains, while developing the questionnaire and

methodology, revealed that the researcher’s note taking and observation of

commuters aroused the interest and suspicion of passengers close by. To

counteract any negative suspicion over the researcher’s activities and

minimise any potential problems, e.g., an observed subject’s perceived threat

of being approached by a stranger on the train; the researcher was clearly

labelled, wearing a badge stating: Researcher.

The length of journey time on each selected route presented another potential

problem. The average journey time into the financial district from the start

point of the observation is 17 minutes8, the second journey away from the

financial centre averaged 26 minutes9. Possible interruptions from embarking

and disembarking passengers, at the frequent station stops, and the

possibility that an informant’s journey might be shorter than the time taken to

conduct the observation and questionnaire were carefully considered.

Consequent piloting and adjusting of the questionnaire’s design and content

ensured that it could be completed with the minimum of inconvenience by the

researcher by circling or marking the informant’s preferred answer. The

format, of researcher completing the questionnaire, limits the inconvenience

factor for the interviewee, which, as noted, is expected to be an issue

affecting data collection. It also enables the most effective use of time and

lessens the possibility of irrelevant data. Average completion time for the

questionnaire is within three minutes.

8 9Average journey times quoted by Transport for London. http://www.tfl.gov.uk/tube/

J M R Quartley 022642 ED3000

9

Tasks The questionnaire10 presents a two-tiered approach. In the first section, self-

reported data on reading practices and Internet usage is collected, from the

observed reader or non-reader. The questions on reading practices were

based upon the same method used in the NALS survey11 and relate to five

print categories: newspapers, magazines, books, personal documents

(referred to in this study as literacy practices at home) and work-related

documents. The questions on Internet use were constructed to appear similar

in style to the reading practices questions. Data is recorded on how often the

subject engages in the different types of reading activities and Internet use, to

produce scores for literacy practices at home (LH), literacy practices at work

(LW) and Internet practices (ICT).

The second section utilises the Recognition Test measures (West, Stanovich

& Mitchell, 1993) and acts as a check against the self-reported data. The tests

developed for this study: Literacy Recognition Test (LRT) and Information and

Computer Technology Recognition Test (ICTRT); use detection logic where

target items (real author names and real Internet related terms) are embedded

among foils (names that are not authors or Internet related terms). The

subject reads the list and selects those names known to be authors or Internet

terms. According to West, Stanovich & Mitchell (1993) there are a number of

advantages to this method:

First the signal detection logic makes the method immune to the social

desirability effects that major contaminants of self-estimates of socially valued

activities such as reading. Guessing is not an advantageous strategy because

it is easily detected and corrected for by examination of the foils checked.

The recognition tests act as a validity check for the self-reported activity

assertions of the Literacy Home/Work and ICT scores.

10 A copy of the questionnaire is included in Appendix 2. 11 NALS survey as described in Smith (1994). See Bibliography/References.

J M R Quartley 022642 ED3000

10

Self-reported measure of reading activities and Internet use

In the first question, subjects are asked to indicate how often they read a

newspaper using the following scale: everyday, few times a week, once a

week, less than once a week, or never. They were also asked to indicate up

to eight sections of the newspaper that they generally read: news, editorials.

Sport, arts/reviews, TV/radio, classified (advertisements), Travel, and

Advice/horoscopes. The second question, asks for the number of magazines

read during the course of a month, on the scale: 0, 1, 2, 3-5, 6 or more. The

third question gathers data for eight categories of books that the subject has

read during the last month: fiction, current affairs, religious/spiritual, history,

reference, manuals, science, and recreation/hobby. Subjects were then asked

(Question 4.) to indicate how often they read a series of different document

types, at home (personal use) and work, using the same frequency scale as

the first question: everyday, few times a week, once a week, less than once a

week, or never. There are six document types: Letters/memos,

journals/magazine articles, manuals or reference, instructions documents,

diagrams/schematics, and bills/invoices.

The next three questions (5, 6 and 7) are concerned with collecting data for

Internet usage. Question 5 asks the frequency of use: everyday, few times a

week, once a week, less than once a week, or never; and in which locations:

home, work, café, or other. Question 6 requests data for the number of hours

per day: less than 1, 1, 2, 3-5, 6-8, 9-10. Question 7 gathers data on which

functions and services the user is engaged: email, search engines, FTP (file

transfer protocol), shopping, web-browsing, news, blogging, e-books, games,

chat, work intranet access, and other.

Recognition Test measures

The last two questions (7 and 8) present the subject with two recognition

tests:

The Literature Recognition Test (LRT). This test comprises of 14 names, 9

authors and 5 foils (these items are listed in Table 8 and 9 in Appendix 1

together with their recognition percentage). In constructing the target list,

J M R Quartley 022642 ED3000

11

names were selected that were considered to be ‘best selling’ authors in a

variety of different fiction genres (e.g. science fiction, travel, romance, fantasy,

crime, etc…). Attention was also taken to include authors that had an

established body of work, over a number of years, to avoid the list being

skewed to a particular age group. The 5 foils were taken from the members of

staff connected to the School of Education at the University of East London.

The 14 names were listed in alphabetical order with the following instructions:

Below you will see an alphabetical list of 12 names. Some of the people in the

list are popular writers and some are not. Read the names and place a mark

next to those you know to be writers.12

The Information and Computer Technology Recognition Test (ICTRT). This

test comprised of 15 names, 10 named terms or acronyms connected with

using the Internet and 5 foils (these items are listed in Table 10 and 11 in

Appendix 1, together with their recognition percentage). In constructing the

target list, names selected were intended to be recognisable to low level users

(e.g. spam, firewall, cookie) as well as higher level users (e.g. Java, Wi-Fi,

phishing). The 5 foil names were taken from other technology areas. The 15

names were listed in alphabetical order with the following instructions: Below

you will see an alphabetical list of 12 names. Some of the people in the list

are popular writers and some are not. Read the names and place a mark next

to those you know to be writers.

The final information gathered related to the subject’s job (to determine their

NRS social grade); age group, using the following categories: 18-25, 26-35,

36-45, 56-65, >6513; and highest academic qualification to date14.

12 The questionnaire wrongly listed the instrument as comprising12 items, there were in fact 14 in the list. 13 The categories as used in West, Stanovich & Mitchell 1993. 14 The list comprised of: GCSE (General Certificate of Secondary Education –end of compulsory education exam for 16 year olds); ‘A’ levels (Advanced Level, a General Certificate of Education usually taken during Further Education and after GCSEs); Vocational qualifications (situated in the work place); Diploma; Degree; Masters; and Doctorate.

J M R Quartley 022642 ED3000

12

Scoring Questions that used the frequency of use answers: everyday, few times a

week, once a week, less than once a week, or never; were scored 5, 4, 3, 2,

1, 0, respectively. For Question 4, the sum of the scores for Home was

divided by 6, being the number of document types, to produce a mean score

for reading literacy practices at home (LH), and similarly for the Work reading

literacy practices (LW). Using this method the mean scores would be in the

range of 0 to 5, where 0 represents no reading activity for all document types,

and 5 represents an everyday reading frequency in all document types.

A mean score was calculated for the sum of Question 5 (scored using the

same method as Question 4) and Question 6. (scored:1 -less than one hour; 2

-one hour; 3 -two hours; 4 –three to five hours; 5 –six or more hours). The

subsequent mean was the measure of Internet use (ICT) and is in the range

0-5, where 0 represents no Internet use, and 5 represents the highest level of

use measurable in the study-everyday for over 6 hours.

The scores are calculated, for the Recognition Test tasks (LRT and ICTRT),

by taking the sum of the selected target items and subtracting the sum of the

selected foils.

The data is categorised into five variables of: Literacy Home (LH) –the sum of

the scores collected for exposure to the different document types read in the

home (Question 4); Literacy Work (LW) –the sum of the scores collected for

exposure to the different document types read at work (Question 4); Internet

use (ICT) –the mean of the values collected for Questions 5 and 6; Literacy

Recognition Test (LRT) –the score for Question 8; and Internet Recognition

Test (ICTRT) –the score for Question 9. The results are variously analysed, to

test the direction and strength of the relationship between the variables

(Spearman's Rank Correlation), and assess whether the means of any two

particular variables are statistically different from each other (t-tests). Further

comparisons on the basis of background variables (age, gender, social class)

are also considered.

J M R Quartley 022642 ED3000

13

Findings There were a total of n=83 subjects observed, of which n=41 were willing to

participate in the survey and this represented a non-participation rate of just

over 50% (proximity to a disembarkation station was the most commonly cited

reason for non-participation). Of these, n=22 were readers (53%) and n=19

non-readers15 (47%). Table 1, below, shows the distribution of age and

educational levels for the readers and non-readers. The largest sample for

both readers and non-readers was from the 36-45 years old group, with non-

reader data almost exclusively concentrated in the age range 26-45 years.

Data collected for education levels was aggregated into three categories;

school; post school/vocational; and tertiary-degree or higher. The reader

group were marginally older (38.4 compared to 35.9 years16) and had a

significantly higher level of education than the non-reader group (2.41

compared to 1.88)17. The highest scores for the variables: LH, LW, and LRT

were displayed by the 46-55 year old age group, with a trend across the

results showing that the older the age group the higher the scores across

these variables (see Table 7 in Appendix 1). ICT scores were broadly similar

across the age range groups. The difference in the mean of the scores for

literacy exposure measures (LH, LW and LRT) and ICTRT is proportionate

across the age range groups, with the exception of 56-65 years group18: 18-

25 Lm=2.86, ICTRT m=3.8 (difference=0.94); 26-35 Lm=3.40, ICTRT m=4.3

(difference=0.90); 36-45 Lm=4.0, ICTRT m=5.2 (difference=1.2); 46-55

Lm=4.26, ICTRT m=5.6 (difference 0.94); 56-65 Lm=3.86, ICTRT m=4.0

(difference 0.14).

15 For the purpose of this study non-reader is classified as not observed reading. 16 This figure is calculated by: selecting the mid-age of each age range category as the age of the individual and dividing by the sample size for reader (n=22) and non-reader (n=19). 17 Education is scored: 1 for school qualification, 2 for vocation/diploma, and 3 for degree. Figures quoted are means. 18 Lm being the mean of the LH, LW and LRT scores listed in Table 7, in Appendix 1.

J M R Quartley 022642 ED3000

14

Table 1 Age and Education distribution

Variable Readers Non-readers Age 18-25 3 1 26-35 4 8 36-45 9 8 46-55 5 1 56-65 1 1 >65 Education (level attained) School qualification 6 6 Vocational/diploma 1 8 Degree 14 4

The data collected did not meet the intended target sample size of six

informants for each grade. The total sample (n=41) was classified as NRS

social grades: A-upper middle class (n=3); B-middle class (n=11), C1 -lower

middle class (n=16), C2 –skilled working class (n=7), D –working class (n=2),

and E –casual/lowest grade workers (n=2). A greater number of samples

(n=30) were gathered on the train journey (J1), into the financial centre, with a

social grade split comprising of: A (n=3), B (n=11), C1 (n=13), and C2 (n=3).

The samples were n=11 for the train journey (J2), out of the financial centre,

with a social grade split of: C1 (n=3), C2 (n=4), D (n=2), and E (n=2).

Analysis of the data from the survey concentrated on the five variables:

Literacy Home (LH); Literacy Work (LW); Internet use (ICT); Literacy

Recognition Test (LRT); and Internet Recognition Test (ICTRT). Results were

tabulated for each informant (see Appendix 1 for individual scores for the five

variables and gender and social grade breakdown) and then the results were

compared at field, social class, gender and reader/non-reader levels, using

Spearman Rank Correlation.

Field level correlations revealed that there was a moderately strong

correlation between LH and LW (r=0.621), which shows some correlation

between the home and work literacy scores. The LRT measure did not

J M R Quartley 022642 ED3000

15

strongly correlate with either the home or work literacy scores, but is more

strongly correlated to home than work scores: LH-LRT (r=0.522) moderate

correlation; and LW–LRT (r=0.394) moderately weak correlation. This is not

unexpected as the LRT measure relates to fiction authors and this genre is

more likely to be encountered in literacy activities taking place in a non-work

setting. A derived variable LH+LW correlates moderately with LRT, relatively

reducing the strength of LH-LRT and supplementing the weakness of LW-

LRT.

The ICTRT measure is not highly correlated with the ICT score (r=0.495,

moderate correlation). There is a moderate correlation for LH-ICT (r=0.459);

weak correlation for LW-ICT (r=0.174); and a moderately weak correlation for

the derived variable LH+LW (r=0.315). The ICT score correlates more

strongly with LH than LW.

The data collected did not meet the intended target sample size of six

informants for each social grade, so the As and Bs were merged (AB n=14)

and C1s and C2s conflated (C, n=28) to create two social groups for

comparison. The means of the AB and C groups were submitted to a series of

two-sample t-tests. There was no significant difference for LH, for either of the

social grades AB and C. However, the results for LW, AB-C were t=3.266,

p=0.0025 [<0.01], 34 df. Relevant means: AB 3.329 (sd 0.350); C 2.826 (sd

0.586). A comparison within the AB group of the means of LH and LW

showed: t=3.401, p=0.0047 [<0.01], 13 df. Relevant means: LH 2.850 (sd

0.407); LW 3.329 (sd 0.350). There was no significant difference for LRT AB

to C, ICT AB to C, and ICTRT AB to C.

Looking at the correlations for the variables of each of the social groups

produced the results set out in Table 2, below.

J M R Quartley 022642 ED3000

16

Table 2 Correlations within conflated social groups: AB; and C Variables AB

r= correlation C r= correlation

LH - LW 0.072 weak 0.739 moderately strong

LH – LRT -0.052 negative 0.593 moderate

LW - LRT 0.424 moderate 0.229 moderately weak

LH+LW - LRT 0.139 weak 0.359 moderately weak

ICT - ICTRT 0.527 moderate 0.596 moderately strong

LH - ICT 0.532 moderate 0.487 moderate

LW – ICT -0.027 negative 0.304 moderately weak

LH+LW - ICT 0.363 moderately weak 0.400 moderate

Males accounted for 65% (n=27) and females 35% (n=14) of the survey.

Among the male group the split between readers and non-readers was 56%

(n=15) to 44% (n=12), respectively. Among the female group the split

between readers and non-readers was equal at 50% (n=7, and n=7,

respectively). Table 3, below, lists the observed reading activities by category

and gender split.

There is a strong correlation for both males and females concerning home

and work literacy: LH-LW male r=0.618; LH-LW female r=0.629 (overall

r=0.621). The literacy activities results displayed noticeably stronger

correlations for males compare with females: LH-LRT: male r=0.556, female

r=0.264; LW-LRT: male r=0.611, female r= -0.079; LH+LW-LRT: male

r=0.593, female r=0.042. Correlation between LW and ICT was noticeably

weaker for males compared to females: LW-ICT male r=0.141, female

r=0.376. The percentage recognition figures for target items and foils in the

LRT and ICTRT measures reveal that females recorded a higher percentage

of foils than males (Tables 8 and 10, respectively in Appendix 1). There is a

significant difference between the mean scores of the male and female

groups in the LRT: male m=11.5 compared with m=6.64 for female. Similarly,

in the ICTRT measure mean scores; male m=9.87 compared with female

m=5.13.

J M R Quartley 022642 ED3000

17

Table 3 Percentage split for male, female and group reading choices.

Reading category Male % Female % Group %

Newspaper 60 - 45

Fiction 13.5 85 36

Non-fiction 6.7 14.5 9

Crosswords 6.7 - 4.5

Adverts 6.7 - 4.5

PDA19 6.7 - 4.5

The mean scores for readers and non-readers for all the variables are set out

in Table 4, below. Group means for LH, LW and LRT are all clearly superior.

The two-sample t-test for these variables show that LH and LRT for the reader

group are significantly higher than the means for the non-reader group (LH

t=3.634, p=0.0009, 35 df; LRT t=4.618, p=0.0001, 34 df) while LW is

approaching significant difference (LW t=2.050, p=0.0502 (ns), 27 df). There

is a similar if more modest, difference between the reader and non-reader

groups for the two-sample t-tests of the ICT and ICTRT variables: ICT

t=3.481, p=0.0014, 33 df; ICTRT t=3.563, p=0.0010, 38 df. The percentage

recognition figures for target items and foils in the LRT and ICTRT measures

reveal that readers have means scores approximately twice the size of non-

readers (LRT males m=11.5, females m=6.6; ICTRT males m=10.2, females

m=5.2), with the two-sample t-tests for these variables: LRT t=4.89,

p=0.00027, 13 df; ICTRT t=4.8, 0.00028, 14 df. In the LRT measure no foils

were recorded for non-readers, while the number recorded for each group in

the ICTRT was similar at Readers n=11, Non-readers n=10.

19 Personal Digital Assistant.

J M R Quartley 022642 ED3000

18

Table 4 Summary report for variables in Reader and Non-reader groups

Variable Readers Non-readers t value

LH 2.94 (sd 0.47) 2.38 (sd 0.54) 3.64 p=0.0009, 35df

LW 3.08 (sd 0.60) 2.53 (sd 1.04) 2.050 p=0.0502, 27df

ICT 3.95 (sd 0.62) 3.16 (sd 0.82) 4.618 p=0.0001, 34df

LRT 7.32 (sd 1.76) 4.42 (sd 2.19) 3.481 p=0.0014, 33df

ICTRT 6.23 (sd 2.94) 3.05 (sd 2.76) 3.563 p=0.0010, 38df

Interpretation

The data collected for readers and non-readers shows that there are sufficient

correlates between variables to suggest that readers, encountered on

commuter trains, will have attained a higher level of education, read more at

home, achieve a higher LRT score and have a higher Internet use than non-

readers. The findings show that differential exposure to print is an indicator for

the levels, and characteristics, of other literacy practices; and supports the

hypothesis that engaging in literacy activities leads to the informal

development of new literacy skills (West, Stanovich & Mitchell, 1993; Smith

MC, 1994).

The data for the social groups revealed that AB s read more at work than C s

(LH AB to C t-test). Conversely, C s read more at home than AB s. The

results of the within group comparison of AB s show that they read less at

home than they do at work. The AB category covers higher and intermediate

managerial, professional, and administrative professions where a large

volume of reading activities in the work place may be assumed as normal.

This higher volume of reading at work probably reduces AB enthusiasm to

partake in reading activities at home. The inference from the data is that the

more an individual reads at work the less likely they are to read at home.

The correlations across gender showed similarly strong results concerning

correlations between home and work literacy. Males were observed reading a

broader spread of document types than females. However, differing scores for

J M R Quartley 022642 ED3000

19

the LRT show a gender skew which favours males. A tentative inference

points to a gender skew in the LRT instrument. This suggests the need for

greater piloting in the development of Recognition Test measures, devised for

this study, so that the instrument works equally well for females as males. At

field level the LRT measure is not informative and is more strongly correlated

with home than work literacy scores. The LRT features popular fiction author

names and is more useful as a broad definition of literacy, such as LH.

Consequently, the weak correlation with LW points to the need for the

development of a Recognition Test instrument to support the more narrow

definition of literacy presented by LW. Although the ICTRT measure did not

suffer from the same flaws, it should also be subjected to more rigorous

piloting prior to use in further studies.

Reading activities are correlated with age and education. There was a

significant relationship between age and literacy activities and Internet usage

scores. Despite the relative similarity in LH, LW, and ICT scores across the

age groups, the performance on the LRT and ICTRT was higher among older

ages, with 46-55 year old scoring highest and the scores dropping down again

for the 56-65 year old group. This inverted-U shaped relationship is similarly

documented in the results of West, Stanovich & Mitchell (1993), where the 40-

50 year old group achieved the highest print exposure scores. The tentative

explanation is that this is driven by the greater levels of print exposure

naturally encountered during the course of a longer life. Overall the results

suggest that age is a good predictor for scores across the five variables.

Similarly, print exposure is an indicator of education level, with readers three

times more likely to have a higher education qualification than non-readers.

The sample is not homogenous. Selecting informants on a commuter train

does not produce a representative spread of data for all background

variables, although this is not entirely unexpected since informants were

selected on the basis of being observed as readers or non-readers. Male

reading practices accounted for two-thirds of the data collected and there is

an unrepresentative spread of data for all the social grades. Random

selection is not a reliable method for achieving the target sample of social

J M R Quartley 022642 ED3000

20

classes at this survey size and this lack of target spread led to analysis being

limited to the conflated groups of AB and C. However, the social class of

subjects encountered, on the two selected journeys, did conform to the

expected social spread outlined in the methodology, with Journey 1 returning

a majority of results for social grades A, B, and C1, while the data for Journey

2 was predominantly for grades C2, D and E. Journey 1 was expected to

present a greater opportunity to collect samples, since the number of

passengers travelling into the city centre was greater than those travelling out,

but the reduced numbers of potential informants on Journey 2 did represent

an obstacle to gathering sufficient data, especially when coupled with a 50%

non-participation rate. Despite the survey’s primary interest in reader and non-

reader data, any future survey for this setting should consider a longer

sampling period to ensure a homogenous sample for gender, social grades

and age group variables, since the limited analysis, in this study, of these

variables highlighted some interesting differences –most notably with AB work

and home reading activities. The questionnaire contained a number of

questions for which data was collected but not used in the analysis (Questions

1, 2, 3 and 7), given the time constraints of the observation setting, for the

administration of the Questionnaire, these could have been omitted without

prejudicing the results of the survey as a whole.

Being a reader suggests an individual is more likely to engage in other literacy

practices, such as Internet use, at a similarly proportionate level. Conversely,

non-readers are non-regular users of the Internet and achieve low scores on

the ICTRT. This points to the reciprocal nature between practice and skills, a

virtuous circle, where the more an individual reads the greater their

accomplishments and proficiency in complex reading tasks, termed “Matthew

effects” by Stanovich (1986) where the ‘rich get richer and poor get poorer’. 20

This study shows that subjects observed as readers significantly out-perform

non-readers across a range of literacy and Internet measures, and that

reading tends to support the notion of a development of cognitive skills. 20 From West, Stanovich & Mitchell (1993)

J M R Quartley 022642 ED3000

21

References

ANGRIST J LAVY V (2001) New Evidence on Classroom Computers and Pupil

Learning. IZA Discussion Paper No. 362 September 2001.

ftp://ftp.iza.org/dps/dp362.pdf

BARTON D & HAMILTON M (1998) Local Literacies: Reading and writing in one

community. London: Routledge.

BELL J (1999) Doing Your Research Project. Maidenhead. Open University

CUNNINGHAM A E AND STANOVICH K E (1998) What Reading Does for The Mind.

American Educator/American Federation Of Teachers, Spring/Summer 1998.

http://www.aft.org/pubs-reports/american_educator/spring_sum98/index.html

GADBERRY, S. (1980). Effects of restricting first graders’ TV-viewing on leisure time use,

IQ change, and cognitive style. Journal of Applied Developmental Psychology, 1, 45–57.

GRADDOL, D., MAYBIN, J. AND STIERER, B. (eds.) 1994 Researching language and

literacy in context, Clevedon, Multilingual Matters.

HAYES & AHRENS (1988) Vocabulary simplification for children: A special case of

‘motherese’? Journal of Child Language, 15.

KERNAGHAN A Literacy as a social practice: a case study. Learning and Skills

Development Agency. Northern Ireland.

http://www.lsda.org.uk/files/lsda/NI/EssentialSkills/Literacy_as_a_social_practice.pdf

MAYBIN, J. (ed.) (1994) Language and literacy in social practice: a reader. Clevedon:

Multilingual Matters with the Open University.

MILROY L (1987) Observing & Analysing Natural Language. Oxford. Blackwell.

MOSER, C.A. AND KALTON, K. (1971) Survey Methods in Social Investigation.

Aldershot, Hants: Gower

NEICE, D.C. (1998), Measures of Participation in the Digital Techno-structure; Internet

Access, ACTS/FAIR Working Paper No.44, Brighton:SPRU, March.

SANKOFF G (1980c) A quantitative paradigm for the study of communicative

competence. In Sankoff G (1980a) The social life of language. Philadelphia: University of

Pennsylvania Press. p47-79.

SAVILLE-TROIKE M (1982) The Ethnography of Communication. Oxford. Blackwell.

SCHON D A, SANYAL B AND MITCHELL W J (1998) High Technology and Low-

Income Communities (Prospects for the Positive Use of Advanced Information

Technology). Cambridge MA : MIT Press.

J M R Quartley 022642 ED3000

22

SMITH M C (1994) What Do Adults Read and Why Does It Matter? Northern Illinois

University. Paper presented at the annual meeting of the Mid-Western Educational

Research Association, Chicago, October 1994.

http://www.cedu.niu.edu/~smith/Unpubs/mwera94_1.pdf

SMITH M C, ELLIOT K S AND HUTCHINSON K M (1994) Do Everyday Reading

Activities Promote Adults’ Cognitive Development. Paper presented at the annual

meeting of the Mid-Western Educational Research Association, Chicago, October 1994.

http://www.cedu.niu.edu/~smith/Unpubs/mwera94_2.pdf

STREET B (1995). Social literacies: Critical approaches to literacy in development,

ethnography, and education. London: Longman.

SUBRAHMANYAM K et al (2001) The impact of computer use on children's and

adolescents' development. Applied Developmental Psychology Vol 22 7±30

http://www.cdmc.ucla.edu/impactcomputer.pdf

TROCHIM W (2002) Research Methods Knowledge Base.

http://www.socialresearchmethods.net/kb/survtype.htm

WEST R F, STANOVICH K E AND MITCHELL H R (1993) Reading in the real world

and its correlates. Reading Research Quarterly, January/February/March 1993

International Reading Association.