Self-Characterization and Attachment Style: A Creative Method of Investigating Children's Construing

This article was downloaded by: [Universita' Milano Bicocca], [Marco Castiglioni]On: 16 May 2014, At: 07:36Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Constructivist PsychologyPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/upcy20

Self-Characterization and AttachmentStyle: A Creative Method of InvestigatingChildren's ConstruingRossella Procacciaa, Guido Veroneseb & Marco Castiglioniba Faculty of Psychology, University eCampus, Novedrate, Italyb Department of Human Sciences “R. Massa,” University of Milano-Bicocca, Milan, ItalyPublished online: 13 May 2014.

To cite this article: Rossella Procaccia, Guido Veronese & Marco Castiglioni (2014) Self-Characterization and Attachment Style: A Creative Method of Investigating Children's Construing,Journal of Constructivist Psychology, 27:3, 174-193, DOI: 10.1080/10720537.2014.904701

To link to this article: http://dx.doi.org/10.1080/10720537.2014.904701

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoever orhowsoever caused arising directly or indirectly in connection with, in relation to or arisingout of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

http://www.tandfonline.com/loi/upcy20

http://www.tandfonline.com/action/showCitFormats?doi=10.1080/10720537.2014.904701

http://dx.doi.org/10.1080/10720537.2014.904701

http://www.tandfonline.com/page/terms-and-conditions

http://www.tandfonline.com/page/terms-and-conditions

JOURNAL OF CONSTRUCTIVIST PSYCHOLOGY, 27(3), 174–193, 2014Copyright C© Taylor & Francis Group, LLCISSN: 1072-0537 print/1521-0650 onlineDOI: 10.1080/10720537.2014.904701

Self-Characterization and Attachment Style: A CreativeMethod of Investigating Children’s Construing

Rossella ProcacciaFaculty of Psychology, University eCampus, Novedrate, Italy

Guido Veronese and Marco CastiglioniDepartment of Human Sciences “R. Massa,” University of Milano-Bicocca, Milan, Italy

This study had the aims of evaluating the self-characterization technique as a research instrumentwith school-age children and investigating whether and how representations of self and others varyas a function of age, gender, and attachment style. Self-characterizations completed by 103 Italianchildren, 47 boys and 56 girls, aged between 7 and 11 years, were analyzed using an ad hoc codinggrid. The Separation Anxiety Test was administered to evaluate children’s attachment style. Agepredicted the process of construing as well as specific content features of the texts, such as level ofabstraction and the type of construct used. In addition, gender was related to the construing processand the expression of emotion, whereas attachment style was associated mainly with stability ofself-image and representation of relationships, and in consequence, the emotional aspects of identity.

Personal construct psychology (PCP) seems particularly suited to the study of child developmentthanks to its own anticipatory perspective that emphasizes change: “Since everyone is ‘developing’or changing from the moment they are born, they can indeed be seen as a form of motion”(Bannister & Fransella, 1980, p. 82). Although scholars within the PCP tradition have madesignificant contributions to developmental psychology (Bell & Bell, 2008; Hayhow, Lansdown,Maddick, & Ravenette, 1988; Mancuso, 2003; Ravenette, 1977, 2003), this field remains lessexplored than others from a personal construct perspective. According to Bannister and Fransella(1980), “it was a great misfortune that Kelly died before he made more explicit his view on theearly development of the construct system” (p. 83).

Given later developments in constructivist clinical psychology (Guidano, 1987; Neimeyer &Mahoney, 1995), we might also argue that, regrettably, Kelly’s premature death prevented himfrom engaging in a direct exchange of views with Bowlby (1969, 1973) in relation to childattachment. It is not by chance that, in reference to the perspective of Salmon (1970), Bannister &Fransella (1980) wrote: “The child’s construing of the mother’s construct system is the jumping-off ground for the development of its own constructing system. Children start out with this anduse it in their dealings with others” (p. 86). Many constructivist psychotherapists assume early

Received 30 March 2013; accepted 3 July 2013.Address correspondence to Marco Castiglioni, Dipartimento di Scienze Umane per la Formazione “R. Massa,” Univer-

sita degli Studi di Milano-Bicocca, P.zza dell’Ateneo Nuovo 1, 20126 Milano, Italy. E-mail: [email protected]

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014

SELF-CHARACTERIZATION AND ATTACHMENT STYLE IN CHILDREN 175

relationships (in particular the relationship with the mother) to play a key role in the developmentof the personal construct system—that is to say, the system of meanings that defines personalidentity (Guidano, 1987; Neimeyer, 2009; Neimeyer & Mahoney, 1995; Procter, 1981; Ugazio,2013; White & Epston, 1990).

From the earliest stages of development, the input received from the family environment isorganized into regular and predictable patterns, forming the basis for the development of twodistinct and complementary processes that become increasingly more complex as the child growsolder: self-perception and perception of the world (Guidano, 1987; Trevarthen & Aitken, 2001).Through sequences of routine actions shared with his or her caregiver(s), the child learns todiscriminate between bipolar dimensions, between familiar patterns (“similar to”) and unfamiliarpatterns (“different from”; Mancuso, 2003). On the basis of caregivers’ responsiveness, thewell-known secure, insecure-anxious-ambivalent, insecure-avoidant, or disorganized styles ofattachment (Ainsworth, Blehar, Waters & Wall, 1978; Bowlby, 1973; Bretherton & Munholland,2008) are developed. Internal working models (IWM), or “interpersonal constructs” regardingthe child’s knowledge of “self-with-other” (Liotti, 1991), are initially developed at a “tacit”procedural level and then elaborated at an “explicit” declarative level (Guidano, 1987), givingrise to “self-narratives” (White & Epston, 1990).

In addition to attachment styles, age appears to play a key role in the development and trans-formation of the child’s construct system. As the child grows older and increasingly experiencesfalsification and invalidation of many of his or her constructs, the way in which he or she con-strues the world increasingly shifts from more concrete, “external,” or “physicalistic” to abstract,psychological, and dispositional constructs (Bannister & Fransella, 1980; Little, 1968).

Gender also appears to significantly influence how individuals construct and communicatetheir images of self and of the world, although it is not yet fully clear what specific differences itdetermines in children’s personal constructs (Bieri, Bradburn, & Galinsky, 1958; Cheshire, 2000;Fivush, Bohanek, Zamanm & Graphin, 2012).

Given this background, the aim of the current study was to explore the dynamics just outlinedas they feature in texts produced by a group of children, using Kelly’s self-characterizationtechnique. Self-characterization represents a typical application of the “credulous” approachpromoted by Kelly, with a specific range of convenience in clinical settings (Androutsopoulou,2001; Armezzani, 2010; Crittenden & Ashkar, 2011; Denicolo, 2003), but less commonly usedin research compared to other means of eliciting personal constructs such as the repertory gridtechnique (Butt, 2008; Fransella; 1995; Fransella, Bell, & Bannister, 2004; Jankowicz, 2004).This is probably due to the greater complexity associated with the textual analysis required bythe self-characterization technique.

Self-characterization is even less commonly applied with children, although some develop-mental studies have reported using it (Jackson, 1988, 1990; Procaccia & Castiglioni, 2010). Afactor limiting its use with children may be the language and cognitive skills required to producea sufficiently complex autobiographical account, in particular the ability to switch from one’sown perspective to that of another person when describing self. This ability, which presumablyrequires possession of a sufficiently advanced theory of mind, develops gradually over time(Fivush & Haden, 2003; Nelson & Fivush, 2004; Procaccia & Di Blasio, 2009) and seems to beconnected with the socialization models to which the child is exposed (Ugazio & Castiglioni,1998). Thus the literature and clinical practice seem to be lacking in research on the validity of the

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014

176 R. PROCACCIA ET AL.

self-characterization method during development, as well as on how children develop personalconstructs and the variables influencing this process.

Therefore, this study was conducted with the aims of (a) evaluating the efficacy of the self-characterization technique as a research instrument with school-age children and (b) investigatingwhether and how representations of self and others vary as a function of age, gender, andattachment style.

METHOD

Participants

The participants in the study were 103 Italian children, 47 boys and 56 girls, aged between 7and 11 years (mean age = 8.84; sd = 1.29), recruited at primary schools in the city of Milanand surrounding area.1 Of middle-class socioeconomic status, 93.2% of the children came fromunited homes, whereas 6.8% had parents who were divorced. The vast majority of the childrenhad one or more siblings (97%). On the basis of their scores on the Separation Anxiety Test (seenext paragraph), 33 children (32%) displayed secure attachment, 37 (36%) anxious-ambivalentinsecure attachment, and 33 (32%) avoidant insecure attachment.

Instruments

The children’s mental representation of attachment style was assessed using the adapted Italianversion (Attili, 2001) of the Separation Anxiety Test (SAT; Klagsbrun & Bowlby, 1976). Thisinstrument evaluates reactions to hypothetical separation from parents, portrayed in a set of sixdigital drawings in which the protagonist is of the same gender as the respondent. Three of thedepicted separations are long (“severe”), and three of short duration (“moderate”).2 By codingthe children’s responses, the following attachment types may be identified: secure, anxious-ambivalent, anxious-avoidant, disorganized/confused.

The children’s personal construct system was accessed by asking them to produce a self-characterization. Kelly’s “classical” instruction was appropriately adapted to match the psycho-logical, cognitive, and linguistic skills of the age group in the sample: “I would like you to talkabout yourself and your personality, as if you were the main character in a story. Write aboutyourself as you would be described by a friend who is fond of you and knows you very well”(Procaccia & Castiglioni, 2010).

Procedure

First, the SAT was individually administered to a larger sample of children in order to selecta research sample (n = 103) containing similar proportions of children with secure, anxious-ambivalent insecure, and avoidant insecure attachment, respectively. The SAT coding was con-ducted by two independent judges (87% agreement), with the intervention of a third judge toresolve cases of disagreement. Only six children from the original sample were found to displaydisorganized attachment, a number too small to be included for the purposes of the current study.The selected children were asked to write a self-characterization during regular class time with

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


their entire class group, in the presence of a teacher and of a researcher who, in line with theresearch protocol, did not intervene in any way. Each child was given the same material (a whitelined sheet of paper) and allowed the same length of time to complete the task (45 minutes).

Analysis of Self-Characterizations

The self-characterization texts were analyzed by two independent judges, with a joint probabilityof agreement of 0.82, using a coding grid (see Table 1) from an earlier pilot study (Procaccia& Castiglioni, 2010) that was developed ad hoc on the basis of the literature (Armezzani,2010; Crittenden & Ashkar, 2012; Fransella & Dalton, 1990; Green, 2004; Jackson 1988, 1990;Landfield, 1971). The grid is divided into four sections covering these areas:

1. Process analysis, focusing on how the narrative was constructed, in terms of sequencesof content and transitions from one theme to another. In particular, this section examinesthe content and emotional valence of the first and last topics mentioned, representing whothe subject is now and what he or she would like to become in the future, respectively(Kelly, 1955); how constructs were developed and arranged in sequence (continuity vs.discontinuity); and the construction of meaning (generalizations vs. specifications). It alsoincludes analysis of the key theme, the level of generality of the constructs (superordinatevs. subordinate), consistency, and length.3

2. Area analysis, involving observation of the constructs making up representations of selfand others, based on Kelly’s hypothesis that “the areas chosen are those in which theclient sees enough uncertainty to make the exploration interesting and enough structureto make it meaningful” (Kelly, 1955, p. 334). In other words, subjects usually speakabout something that is meaningful for them but yet deserving of further elaboration.Specifically, this section of the grid records the quantity and nature of references toself and others, assessing the level of abstraction of the representations by measuringuse of concrete and external constructs versus abstract, psychological, and internal ones(Bannister & Fransella, 1980; Salmon, 1970). With regard to the representation of others,area analysis also includes observation of the figures—both internal and external to thefamily—around which subjects organize their constructs. In Kelly’s view, the subjectschoose contextual areas because they identify particularly strongly with them, a preferencethat also serves to distinguish them from other people. Finally, this section also recordsdifficulties regarding subjects’ self image or image of others, via indicators of syntacticor grammatical disruptions of the text, such as abrupt changes of topic (avoidance) orinterruptions in the flow of the narrative (unfinished sentences).

3. Thematic analysis, involving counting the prepositions, conjunctions, and adverbialphrases expressing relationships of causality among the themes discussed (since, giventhat, because of, on account of , etc.), while also recording the dominant type of locus ofcontrol (internal vs. external). This section sheds light not only on what is important tothe subject but also on how they experience their life contexts (Armezzani, 2010).

4. Dimensional analysis, examining the dimensions of meaning, in terms of the constructssubjects use to construe their experience. It includes observation of the time dimension(past, present, future) “to understand the lay of those channels through which the client’spsychological processes flow in search of the future” (Kelly, 1955, p. 337). Dimensional

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


TABLE 1Content Analysis of the Self-Characterizations: Coding Categories Divided into Four Sections (Process,

Contextual Areas, Thematic Analysis, and Dimensional Analysis)

Section Coding categories

Process Sequences and transitionsAnalysis of the first topic (reference to self vs. reference to others)Emotional valence of the first topic (positive vs. negative vs. neutral references)Analysis of the final topic (reference to self vs. reference to others)Emotional valence of the final topic (positive vs. negative vs. neutral references)Type of elaboration: analysis of transitions from one sequence to the next (continuity vs.

discontinuity)Narrative mode (generalization vs. specifications)

Organization: Analysis of the core themeDegree of generality (superordinate vs. subordinate constructs)Consistency (instances of contradictions)Length of texts (n. of words)

Contextual areas Representation of othersIdentification of other people (n. of references to mother, father, brothers/sisters, other

family members; friends or other significant figures)Analysis of types of reference (personal appearance vs. abilities, activities and

preferences vs. personality and general statements)Affective valence of references, i.e., esteem for/trust in others (positive vs. negative)Relationship difficulties (instances of avoidance; unfinished sentences)

Representation of selfAnalysis of types of reference (personal appearance vs. abilities, activities and

preferences vs. personality traits and general statements)Affective valence of references, i.e., self-esteem/self-confidence (positive vs. negative)Difficulty in representing self (instances of avoidance; unfinished sentences)Analysis of emotional aspects in relation to self and others (references to joy/happiness,

sadness, anger, fear, guilt, shame)Thematic analysis Cause-effect relationships

Analysis of the locus of control (reference to internal vs. external causes)Type of external causes (events vs. people)

Dimensional analisys Temporal dimensionReferences to past, present, and futureAffective valence of references (positive vs. negative)

Level of abstraction: Action vs. thoughtReference to actionsReflective abilities: the mental or internal states of self or others

Type of constructReferences to similarities (like, similar to, etc.) vs. references to contrasts (in contrast,

on the other hand, etc.)

analysis also includes assessment of the level of abstraction of the elicited constructs (thedimension of action vs. the dimension of thought) and of type of construct in terms ofthe dominance of particular similarities (terms such as like that, in the same way, etc.) orcontrasts (on the contrary, on the other hand, etc.), in order to identify the dichotomousalternatives between which subjects find themselves continually having to choose.

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


Atlas-Ti software for textual analysis was used to process the data. This instrument, designedto facilitate qualitative content analysis, is based on a network model that enables data to beorganized and the results synthesized on the basis of guidelines defined by the researcher. Codingmay be automatic or manual. We chose the latter option, which, although it provides for reiteratedcomparison procedures and systematic interpretation of the textual data, is similar to a “paperand pencil” coding procedure, albeit faster and more flexible. It therefore represents a goodcompromise between a coding system that remains closer to the text and one that is more abstractbut with scope for synthesizing the data (Veronese, Procaccia, Romaioli, Barola, & Castiglioni,2013; Veronese, Romaioli, & Castiglioni, 2012). The analysis procedure involved importingthe texts into the software and carefully observing them in order to select words, phrases, orparagraphs to link to the mutually exclusive categories listed in the grid above.

Quantitative analysis was then carried out using SPSS. Specifically, the raw data were trans-formed into percentages on the basis of the number of words produced by each child, andKendall’s tau-b index was calculated to check for correlations between age, attachment style,and construct system. Finally, in order to obtain a further indication of how age, gender, andattachment style individually influenced the contents of the self-characterizations, a series ofblock multiple regression analyses were conducted.

RESULTS

The Feasibility of Self-Characterization Technique in Childhood

The qualitative analysis of the texts confirmed that the self-characterization technique is aneffective instrument for use with children (see Table 2). With regard to the construing process,the children in the sample demonstrated the ability to produce self-narratives of sufficient length,consistency, and detail. The themes were linked together in coherent sequences, with an overallhigh degree of continuity between the constructs expressed. With regard to the contextual areas,the children’s representations of self and others could be inferred from their narratives. Thefigures most frequently referred to were friends, and in the family context, mother, father, andsiblings, in that order. As was to be expected given the age of the participants, descriptions of selfand others were mainly composed of concrete constructs, regarding the activities and abilitiesof self and others, followed by physical appearance (physical and situational constructs of the“external” kind), with a smaller proportion of references to personality traits (“psychological” and“personality” constructs). However, the children made frequent reference to their own emotions,both positive (joy) and negative (guilt, anger, fear, and sadness). To a lesser extent, their accountsalso contained references to the emotions of others: In this regard, self-characterizations alsoappear to be a useful instrument for the identification of dominant emotions. With regard to thedimension of time, these school-age children appeared to be strongly anchored to the present,but were also able to give an account of their past histories and, to a lesser extent, to discusstheir futures. Concerning locus of control, they appealed more frequently to external causes,particularly events, making less reference to internal factors. They constructed reality mainly interms of the dimension of action, although they also alluded to thought and reflection. Finally,the dimensional analysis shed light on the dimensions of meaning or bipolar constructs used bythe children to interpret their experiences, expressed in terms of similarity and contrast.

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014

TAB

LE2

The

Fea

sibi

lity

ofS

elf-

Cha

ract

eriz

atio

nTe

chni

que:

Des

crip

tive

Sta

tistic

sof

the

Ana

lysi

sof

Text

s(P

erce

ntag

esfo

rN

omin

alV

aria

bles

;Mea

nan

dsd

for

Num

eric

alV

aria

bles

)

Sect

ion

Cod

ing

cate

gori

es

Proc

ess

Sequ

ence

san

dtr

ansi

tion

sFi

rstt

opic

Ref

eren

ces

to:

Em

otio

nalv

alen

ce:

self

=94

.2%

Posi

tive

=32

%ot

hers

=5.

8%N

egat

ive

=9.

7%N

eutr

al=

58.3

%L

astt

opic

Ref

eren

ces

to:

Em

otio

nalv

alen

ce:

self

=72

.8%

Posi

tive

=47

.6%

othe

rs=

27.2

%N

egat

ive

=21

.4%

Neu

tral

=31

.1%

Type

ofel

abor

atio

nC

ontin

uity

16.5

4(7

.87)

Dis

cont

inui

ty8.

13(4

.98)

Nar

rativ

em

ode

Gen

eral

izat

ions

13.7

8(6.

80)

Spec

ifica

tions

11.6

4(7

.32)

Org

aniz

atio

n:an

alys

isof

the

core

them

eD

egre

eof

gene

ralit

ySu

pero

rdin

ate

cons

truc

ts12

.52

(7.1

1)

Subo

rdin

ate

cons

truc

ts6.

86(5

.77)

Con

sist

ency

(N.o

fco

ntra

dict

ions

)1.

03(5

.08)

Len

gth

(N.o

fw

ords

)99

.92

(57.

64)

Con

text

uala

reas

Rep

rese

ntat

ion

ofot

hers

mot

her

0.7

(0.9

5)fa

ther

0.6

(0.9

)si

blin

gs0.

53(0

.85)

othe

rfa

mily

mem

bers

0.44

(0.9

9)fr

iend

s1.

40(2

.65)

othe

rs0.

53(1

.04)

Type

ofre

fere

nce

Pers

onal

appe

aran

ce0.

32(0

.8)

Abi

litie

s/ac

tiviti

es0.

90(1

.82)

Pers

onal

ity0.

26(0

.78)

Val

ence

ofre

fere

nces

Posi

tive

refe

renc

es0.

26(0

.78)

Neg

ativ

ere

fere

nces

0.32

(0.6

4)R

elat

ions

hip

diffi

culti

esIn

stan

ces

ofav

oida

nce

0.62

(1.8

0)U

nfini

shed

sent

ence

s0.

08(0

.37)

180

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014

Rep

rese

ntat

ion

ofSe

lfTy

peof

refe

renc

ePe

rson

alap

pear

ance

3.76

(4.4

2)A

bilit

ies/

activ

ities

4.86

(4.0

1)Pe

rson

ality

1.56

(2.1

8)

Aff

ectiv

eva

lenc

eof

refe

renc

esPo

sitiv

ere

fere

nces

1.11

(1.9

8)N

egat

ive

refe

renc

es1.

03(1

.60)

Dif

ficul

ties

inre

pres

entin

gSe

lfIn

stan

ces

ofav

oida

nce

1.99

(4.6

0)U

nfini

shed

sent

ence

s0.

78(1

.30)

Em

otio

nsre

gard

ing

self

Joy

0.40

(0.7

2)Sa

dnes

s0.

18(0

.42)

Ang

er0.

29(0

.75)

Fear

0.17

(0.6

9)G

uilt

0.29

(0.7

2)Sh

ame

0.02

(0.1

5)E

mot

ions

rega

rdin

got

hers

Joy

0.12

(0.1

5)Sa

dnes

s0.

03(0

.19)

Ang

er0.

08(0

.03)

Fear

0.06

(0.2

5)G

uilt

0.04

(0.2

0)Sh

ame

0(0

)T

hem

atic

anal

ysis

Cau

se-e

ffect

rela

tion

ship

Inte

rnal

caus

es0.

30(0

.83)

Ext

erna

lcau

ses

(eve

nts)

1.42

(1.1

6)E

xter

nalc

ause

s(p

eopl

e)0.

28(0

.66)

Dim

ensi

onal

anal

ysis

Tem

pora

ldim

ensi

onN

.ref

eren

ces

topa

st1.

65(2

.92)

N.r

efer

ence

sto

pres

ent

9.38

(5.9

0)N

.ref

eren

ces

tofu

ture

0.60

(1)

Val

ence

ofre

fere

nces

Posi

tive

0.41

(0.8

3)N

egat

ive

0.8

(0.3

7)L

evel

ofab

stra

ctio

nC

oncr

ete

cons

truc

tsba

sed

onac

tions

6.58

(10.

35)

Con

stru

cts

rega

rdin

gre

flect

ive

capa

city

:M

enta

lsta

tes

ofot

hers

0.24

(0.6

2)M

enta

lsta

tes

ofse

lf3.

11(3

.08)

Type

ofco

nstr

uct

Sim

ilari

tyco

nstr

ucts

0.56

(1.0

5)C

ontr

astc

onst

ruct

s1.

03(5

.08)

181

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


Correlations Between Age, Gender, and Attachment Style in Self-Characterization

The results of the correlational analysis (Table 3) conducted using Kendall’s tau-b index, broadlyconfirmed the data observed in the literature reviewed above. To illustrate the statistical relation-ships described below, a pair of sample self-characterizations are provided for each of the threevariables under study.

Age

The values obtained indicated that age was associated with a number of aspects of self-characterization. To illustrate these statistical relationships, we provide the self-characterizationsof two children, aged 7 and 10 years, respectively4:

Mary, age 7 years:

I have a dog called Billy. I’m seven. I have brown hair and brown eyes too. I like playing tennis,even though I don’t score many points. I am not very tall and I’m a bit thin too. I like math. My bestfriend’s name is Giovanni and he has blond hair and blue eyes. I get good marks at schools, but I getreprimands and bad marks too. I get them in geography because the teacher is strict. Bye.

Laura, age 10 years:

My name is Laura and I am ten. I am a bit plump but I am tall. I have brown hair and blue eyes. Ihave a kind and sensitive personality, although I am a bit sulky.I attend primary school in X. I live in X in X Street. I love pasta with tomato sauce and meat, but Ihate fish. I like watching TV, especially cartoons and musicals, because I like singing, but I don’t likewar films because I hate violence . . .I like playing volleyball and my coach says I’m good at it. My best friend’s name is Maria. I like hera lot because she is gentle and always cheerful.I am doing well at school, my favorite subjects are history and geography, but I don’t like math. Mylowest marks are in math because I don’t study it as much, when I get a bad mark I am sad andsometimes I cry.I am a sociable and gentle girl, although I sometimes get mad with my friends and we have anargument. But then I feel bad, I get sad and so I say sorry and we make friends again. My Mum’sname is Anna, my Dad’s name is Marco and my little bother is called Lorenzo, he’s seven. I get onwell with Andrea too, he is cheerful and funny although sometimes he makes me mad.When I was smaller I used to like drawing, now I prefer reading, especially books about elves andfairies.When I’m big I would like to be a journalist like my Dad, but maybe I won’t be as good as he is.

Specifically, with regard to the process through which the children constructed their imagesof self and others, age was negatively correlated with the content and emotional valence ofthe first topic and with the level of contradiction. In contrast, it was positively correlated withlevel of continuity, specifications, relationships between constructs, and length of narrative. Thismeans that as age increases, the ability to focus on the self as the protagonist of life eventsbecomes more advanced, the construct system becomes more complex, and representations morepositive. The children become increasingly competent in terms of producing long, detailed, andcoherent narratives that link topics at different levels of abstraction, indicating a higher degree ofelaboration. Similarly, with regard to the contextual areas elicited, age was positively correlated

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


TABLE 2Correlations Between Age, Gender, and Attachment Style in Self-Characterizations

Coding categories Age Gender Attachment

ProcessContent first topic −0.220∗∗ — —Emotional valence first topic −0.169∗ −0.312∗ 0.335∗∗

Emotional valence last topic — — 0.202∗

Continuity 0.224∗∗ — −0.171∗

Generalizations 0.144∗ −0.167∗ —Specifications 0.155∗ — —Superordinate constructs 0.160∗ — —Subordinate constructs 0.319∗∗ — —Contradictions −0.170∗ −0.271∗∗ 0.271∗∗

Words 0.220∗∗ 0.240∗∗ −0.249∗∗

Contextual areasMother — — −0.315∗∗

Father — — −0.379∗∗

Siblings — — −0.292∗∗

Friends — 0.191∗ −0.217∗∗

Others — — −0.351∗∗

References to the abilities/activitiesof others

0.262∗∗ — —

Positive references to others — 0.265∗∗ –0.295∗∗

References to the abilities/activitiesof self

0.173∗ — —

Avoidance (others) — — 0.188∗

Unfinished sentences (others) — — 0.194∗

References to personality-self 0.148∗ 0.260∗∗ –0.203∗∗

Positive references to self 0.152∗ — –0.241∗∗

Avoidance (self) — –0.404∗∗ 0.515∗∗

Unfinished sentences (self) — –0.193∗ 0.242∗∗

Joy — 0.328∗∗ −0.410∗∗

Sadness 0.172∗ — —Guilt 0.186∗ — —Total emotions regarding self 0.161∗ 0.273∗∗ –0.232∗∗

Joy others — 0.285∗∗ —Anger others — — 0.267∗∗

Total emotions regarding other 0.196∗ 0.196∗ —Total emotions 0.200∗∗ 0.273∗∗ –0.161∗

Thematic analysisInternal causes 0.179∗ 0.186∗ –0.272∗∗

External causes–events 0.205∗ — —External causes–people 0.159∗ — —Total external causes 0.275∗∗ — 0.189∗

Dimensional analysisReferences to present — 0.182∗ —References to past — −0.202∗∗ —References to future 0.172∗ — −0.272∗∗

Positive references to future −0.247∗∗ — −0.256∗∗

Negative references to future — — 0.250∗∗

Action — −0.296∗∗ 0.286∗∗

Mental states of self 0.288∗∗ 0.296∗∗ −0.390∗∗

Mental states of others 0.189∗ 0.285∗∗ −0.267∗∗

Total mental states 0.189∗ 0.301∗∗ −0.419∗∗

Similarities 0.299∗∗ — —Contrasts 0.199∗ — −0.244∗∗

Note: ∗p < .05; ∗∗p < .01.

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


with references to the activities and abilities of self and others, references to the narrator’sown personality, and level of positive self-reference. Age was also positively correlated withreferences to sadness and guilt and with the overall percentage of emotions reported. Therefore,older children tend to construct images of themselves and of others at a higher level of abstraction,increasingly adding “contextualized” and “psychological” constructs to the more physical andconcrete constructs favored by the younger children (Bannister & Fransella, 1970; Salmon,1970).

Older children also appear to make more frequent self-references, implying a gradual strength-ening of self-esteem and self-confidence. In addition, they display gains in the ability to expressemotion, progressing from a focus on external experiences to a greater emphasis on the internalrepresentations of self and others. They show particular gains in awareness of emotional am-bivalence (Feixas & Villegas, 1991; Harris, 1989; Ugazio & Castiglioni, 1998) and expressionof negative emotions, such as sadness and guilt, indicating a more advanced ability to elaboratethe events and relationships in which they are involved: When the latter conflict with children’sconstruct systems, this gives rise to negative emotions and feelings of inadequacy. Similar re-sults emerged from the thematic analysis: Age was positively correlated with causal attribution,both internal and external. This means that as children grow older, their increasing cognitiveand affective maturity leads them to use fewer concrete constructs and a higher proportion ofevaluative constructs: The latter are connected both with the attempt to attribute a deeper levelof meaning to experience and with greater reference to the social context. Finally, with regardto the dimensional analysis, a positive correlation was found between age, on the one hand, andchildren’s references to the future, their own mental states, similarities, and contrasts, on the other.However, age was negatively correlated with negative references to the future, implying greaterability on the part of older children to plan for the future, albeit with a corresponding increase infear and doubt. Children’s reflective capacity also increases with age, as borne out by the olderparticipants’ greater focus on their own mental states and superior ability to compare themselvesto others in search of both shared traits (similarity constructs) and the distinctive qualities of theself (contrast constructs).

Gender

Gender was also found to be correlated with the constructs through which the childrenrepresented the world and themselves. Again, we provide by way of illustration the self-characterizations of two 9-year-old children, one male and one female.

Luca, male, age 9 years:

Hi. My name is Luca and I am nine years old. I have black hair and brown eyes. I like playing footballand computer games. I like watching TV, especially “Zac & Cody.” And now I can’t think of anythingelse. . . . Ah yes, I can add that my favorite color is orange and that I am an A.C. Milan supporter. Ilove dogs, but my sister Laura doesn’t. I am good at school and I get good marks, but sometimes theteacher gets mad at me and she scolds me because she says I talk too much, but really it’s becausemy classmates disturb me and I turn around to tell them to stop it.

Alessandra, female, age 9 years:

Hi. My name is Alessandra, I am nine years old. I have nice blond hair and brown eyes and I like towear a hairband.

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


I am a lively girl, I like to laugh and spend time with my [girl] friends.I am also tidy and I often help my Mum in the kitchen, because I enjoy that a lot. I have a big sistertoo called Francesca who is thirteen. I like playing with her, although now she sometimes prefers togo out with her friends.My greatest desire is to become a vet because I love animals. At the moment my family has a turtle,a goldfish, and a cat called Max.I’m very good at skiing and last year I even won a medal because I came first in a competition.Thank you, I’ve finished now.

With regard to the process of construction, there was a negative correlation between gender,on the one hand, and emotional valence of the first topic, number of generalizations, number ofwords, and contradictions, on the other (see Table 3). This implies that girls tend to produce longerself-characterizations that are more detailed, more coherent, and more focused on the self. Withregard to contextual areas, gender was positively correlated with references to friends, positivereferences to others, and positive references to the narrator’s own personality. In contrast, it wasnegatively correlated with instances of avoidance and unfinished sentences. In other words, girlsmade more frequent mention of friends and in general were more inclined to refer to others inpositive terms. With regard to the constructs used to represent the self, girls appeared to havegreater recourse than boys to abstract and psychological constructs. In addition, girls displayedless difficulty in speaking about themselves, as borne out by the negative correlation betweengender and instances of avoidance and unfinished sentences.

Concerning the emotional dimension, the correlational analysis suggests that girls find iteasier to focus on their own emotions and on those of others, particularly on positive emotionssuch as joy. The thematic analysis showed a positive correlation between gender and internalcausality: Girls made more frequent reference than boys to an internal locus of control, perceivingthemselves as competent in the management and control of events.

Finally, the dimensional analysis yielded negative correlations between gender and referencesto the past, suggesting that boys tend to remain more attached to the dimensions of the past thangirls. The constructs drawn on to represent the world and its meanings also varied as a function ofgender: Boys were more inclined to use the dimension of action and girls the dimension of thought,as borne out by the greater number of references to mental states in the self-characterizations ofthe girls.

Attachment

The following examples are drawn from the self-characterizations of Sara, with a secureattachment style, and Sonia, with insecure-avoidant attachment, both aged 10.

Sara, secure attachment:

Hi, my name is Sara. I live in Milan with my family: My Mum is called Laura, my Dad Vincenzoand my little brother who is five years old is called Luca. I also have a cat named Birillo and maybefor my birthday my parents are going to give me a dog too and I am very happy about that and Ican’t wait for it to arrive. I have long brown hair; I’m quite tall and thin. I love to spend time withmy best friends Paola and Marta, telling one another dreams, secrets and jokes. I have a great timewith them and I’m happy when my Mum lets me invite them to my house after school. I like doingdancing and gymnastics. My dance teacher is very nice; her name is Elisa and she says that I amvery good. My favorite subject is science but I like drawing too and my biggest dream for when Igrow up is to be a fashion designer and design clothes. I get good marks at school because I like

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


studying, especially science and Italian. My favorite food is pizza, but also pasta with the sauce myMum makes, because she’s a very good cook. I like playing with my brother, especially at GuessWho, although we sometimes fight about who gets to go first. But then we soon make friends againand we forget all about it! He likes playing football, but I prefer volleyball. I like to dress smartly andwear something different everyday. That’s why my biggest dream is to be a fashion designer! I lovegoing to the sea and also to the mountains where I go for wonderful walks with my Dad who knowsall the paths. I love gathering mushrooms with him because he is an expert and he tells me whichones I can pick and which ones I cannot. I am quite playful, although sometimes I’m a bit stubborn,but most of the time I know how to be a true friend!

Sonia, insecure-avoidant attachment:

Hi, my name is Sonia and I’m ten years old. I have brown hair and brown eyes. I am average heightand thin.I like dancing and watching TV. I’m not so good at school, I don’t like math and geography. SometimesI get bad marks because the tests are too difficult. I like reading but not so much. I have a six-year-oldbrother, Pietro, who always plays mean tricks on me and then I get mad and I scold him.

The analyses regarding process showed a positive correlation between attachment style, onthe one hand, and emotional valence of the first and last topics and level of contradictions, onthe other. In contrast, a negative correlation was found between attachment style and level ofcontinuity and number of words. This implies that insecure attachment styles, particularly theinsecure avoidant pattern, are associated with shorter accounts, and poorer and less accessiblerepresentations of self and others, with a greater focus on others, although often with negativeconnotations. The narratives of insecurely attached children were characterized by less detailand elaboration and more frequent contradictions, perhaps a reflection of their defensive need todistance themselves from critical themes that could give rise to negative emotion.

Similarly, the contextual area analysis showed a negative correlation between attachmentstyle and references to parents, siblings, friends, and other people in general, whereas a positivecorrelation was found between attachment and instances of avoidance and unfinished sentencesin the descriptions of self and others, and references to anger experienced by others. This suggeststhat insecurely attached children tend to exclude others from their representations of self, bothwithin and outside of the family context, and when they do refer to others, they represent themmore negatively, perceiving them as hostile. They also tend to construct a more negative self-image, use fewer psychological constructs than their securely attached peers and, as reflectedin the greater incidence of avoidance and unfinished sentences, appear to have a less stable andmore fragile sense of self and greater relationship difficulties.

Turning to emotion, insecure attachment is linked to greater difficulty in focusing on thefeelings of self and others, particularly on positive emotion. Concerning thematic analysis, thenegative correlation between attachment and appeal to internal causes and positive correlationbetween attachment and appeal to external factors, indicate that insecurely attached children morefrequently identify the locus of control as external, probably because they perceive themselvesas incompetent and unable to manage the situations they encounter.

Finally, with regard to the dimensional analysis, insecurity was positively correlated with agreater focus on the present, probably due to difficulty in planning for the future, as suggested bythe lower percentage of references to the future. When the future was mentioned by insecurelyattached children, it tended to be negatively described. In sum, insecurely attached children appearto construe reality predominantly in terms of concrete features, linked to action, whereas they are

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


poorer than their securely attached peers at engaging in reflective thinking aimed at understandingthe mental states of self and others.

Regression Analyses

In order to tease out the distinct and specific influence of each of the variables age, gender,and attachment style on the personal construct system, we conducted block multiple regressionanalyses, with the various coding categories entered as criterion variables and age as predictorin the first block, gender in the second, and attachment in the third (recoded as a dichotomousvariable: secure-insecure; see Table 4).

Age was found to predict type of construct used (particularly the degree of continuity), levelof elaboration (generalizations vs. specifications), and complexity of the constructive system(superordinate vs. subordinate constructs). In relation to content, age predicted the level ofabstraction with which children represented themselves and others (activities/preferences) aswell as their reflective capacity (references to the mental states of others). It also predictedreferences to locus of control, both internal and external and, at the emotional level, the incidenceof guilt, a complex emotion that demands more advanced social understanding. With regard tothe dimensions of meaning, age predicted references to the future, negative valence, and type ofconstructs used (similarities).

Gender predicted the degree of elaboration of constructs (generalizations vs. specifica-tions), their degree of complexity (superordinates vs. subordinates), and length of the self-characterization narrative. In terms of content, gender predicted the level of abstraction withwhich children represented themselves (references to personality) and difficulties in representingthe self (avoidance). Gender also predicted overall references to emotion, in particular joy, theprevalent temporal dimension cited (past), and type of representation (action-based).

Attachment style appeared to have a lesser influence on the process of construing than age, inthis area only predicting the coherence of the narrative (contradictions). In contrast, it seemed toexercise a particularly broad influence on the contents of the representations of self and others.This was especially the case with regard to ability to refer to self (as may be inferred from thelevel of avoidance and unfinished sentences) or to significant others (see references to mother,father, other figures: level of avoidance and unfinished sentences) and the emotional valence ofreferences (negative vs. positive). With regard to emotion, attachment predicted positive emotions(joy) and also the perception of hostility on the part of others. It also predicted causal attribu-tion, both internal and external. In addition, attachment style predicted the temporal dimensions(present) and emotional valence (positive references to the future). Finally, it predicted the levelof abstraction of constructs—that is to say, the characteristic ways in which children construe andinterpret reality—as well as the presence in their self-representation of the dimensions of actionand of thought (reflective capacity).

DISCUSSION

The findings of this study are substantially in line with those of earlier research (Jackson, 1988,1990; Procaccia & Castiglioni, 2010). This seems to confirm that self-characterization techniqueis a valid instrument for use with children. Age and degree of maturity are associated with the

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


TABLE 3Multiple Regression Analyses (Coding Categories as “Criterion Variables”; Age, Gender, Attachment Style as

“Predictors”)

Criterion Predictors β t Sig. R-square

Type of elaboration Continuity Age .342 3.599 .001 .141Narrative mode Generalizations Age .293 3.060 .003 .129

Gender −.277 −2.726 .008Specifications Age .216 2.096 .039 .080

Level of generality Superordinate constructs Age .267 2.760 .007 .107Gender −.259 −2.524 .013

Subordinate constructs Age .425 4.570 .000 .177Consistency Contradictions Attachment .251 2.559 .012 .165Length N. Words Gender .187 1.806 .05 .095Representations of others mother Attachment −.221 −2.122 .036 .059

father Attachment −.328 −3.243 .002 .110others Attachment −.258 −2.538 .013 .102Positive ref. to others Attachment −.213 −2.055 .042 .069Abilities/activities Age .281 2.895 .005 .106

Relationship difficulties Avoidance Attachment .209 2.058 .042 .102Representation of self Personality Gender .218 2.091 .039 .081Affective valence Positive ref. to self Attachment −.420 −4.273 .000 .159Difficulties in representing self Avoidance Gender −.237 −2.411 .018 .185

Attachment .218 2.253 .026Unfinished sentences Attachment .267 2.664 .009 .126

Emotions concerning self Joy Gender .226 2.369 .020 .232Guilt Attachment −.342 −3.632 .000

Age .190 1.924 .057 .072Emotions concerning others Joy Gender .250 2.419 .017 .097

Anger Attachment .208 2.010 .047 .072Total emotions mentioned Total emotions Gender .208 1.993 .049 .082Locus of control Internal causes Age .255 2.800 .006 .209

Attachment −.339 −3.553 .001External causes Age .361 3.967 .000 .210

Attachment .220 2.306 .023Temporal dimension Past Gender −.269 −2.621 .010 .110

Present Attachment .204 2.065 .042 .075Positive ref. to future Attachment −.285 −2.774 .007 .080Negative ref. to future Age .280 2.880 .005 .099

Level of abstraction Ref. to actions Gender −.240 −2.313 .023 .091Reflective abilities Mental states of self Attachment −.377 −3.958 .000 .212

Mental states of others Age .355 3.904 .000 .214Attachment −.236 −2.477 .015

Total mental states Attachment −.406 −4.305 .000 .226Type of construct Similarities Age .290 2.935 .004 .092

Note: Significativity fixed at p < .05, p < .01 level.

ability to fully express one’s construct system through the verbal channel: As age increases,so does ability to elaborate meaning in an ever more complex and coherent way, as previouslyreported in the literature (Bannister & Fransella, 1980; Fivush, 2011; Salmon, 1970). Kelly(1955) claimed that the construct system changes as a function of how the individual anticipates

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


events. These changes may be slight and only affect the subordinate constructs, or they may bedeeper and affect core constructs, when the latter have proven inadequate in relation to events. Inorder to address difficulties encountered in anticipating events, as children grow older they applyprogressively more subtle discriminations: Constructs become more differentiated and abstract,with the emphasis shifting from physical and situational aspects to a representation that is morefocused on the psychological dimension (Little, 1968).

Emotional aspects also become more important, hand in hand with an enhanced capacity toattend to the internal states of the self and others. Increased maturity is also accompanied by theelaboration of more complex forms of relations between constructs, in terms of interconnections(Procaccia & Castiglioni, 2010). Social comparison plays an increasingly strong role in theconstruction of meaning (Ugazio & Castiglioni, 1998), linked to the requirement to understandone’s own uniqueness, but at the same time to identify similarities with others, thus satisfying theneed to belong to a group (Leung, Marsh, Craven,Yeung, &Abduljabbar, 2013).

Gender also influences both the way in which children attribute meaning to the self and tothe world around them and the specific contents of these meanings. In line with the findings ofearlier studies on gender differences in child development (Fivush et al., 2012), our data confirmthe tendency of girls to produce longer autobiographical accounts that are more coherent anddetailed, present a greater degree of reflection and evaluation, and are characterized by morefrequent recourse—compared to male peers—to abstract, internal, and psychological constructs.Girls also organize experience predominantly on the basis of emotional constructs, with a lesseremphasis on the dimension of action, which in contrast plays a central part in the construingof boys. Similar differences have been found in studies with adults (Zaman & Fivush, 2013a),in which mothers displayed a greater tendency than fathers to adopt elaborative conversationalstyles, especially with daughters, speaking in more detailed terms about the past and aboutemotions, involving their children in conversation through the use of open questions and helpingthem to construe meanings and develop their own narrative abilities.

Turning to attachment styles, this factor has a strong influence on how children differentiatethe polarities of “self” and “nonself” (Mancuso, 2003), and on how they differentiate and extendtheir entire personal construct system. Similarly to what has been found in relation to adults(Main 1991; Parkes, Stevenson-Hinde, & Marris, 1991; Singer, Blagov, Berry & Oost, 2013;Zaman & Fivush, 2013b), insecure children find it difficult to provide a detailed, coherent, andpositive account of themselves and others, manifesting both personal and relationship difficulties.In particular, insecurity leads children to avoid engaging with their caregivers and with otherpeople in general, and when this does happen, their representation of the relationship remains ata more superficial level, based on “external” constructs. Insecure children, on the basis of theirexperience that caregivers are either not emotionally available or only intermittently so, havegreater difficulty in recognizing and expressing their own emotions and those of others (Fonagy& Target, 1997). They therefore tend to develop a construct system that is less abstract than that ofsecurely attached children, based on the dimension of action more so than on reflective abilities, inline with the findings of studies on attachment that have reported more advanced metacognitiveawareness in secure children: Insecurity is linked with a lesser inclination to reflect on one’sown experience (Fonagy & Target, 1997; Wallin, 2007), giving rise to difficulty in creatingpropositional representations and drawing on metarepresentations (Main, 1991).

In sum, age, gender, and attachment style may all be said to have a significant impact on thedevelopment of personal constructs in children. If, however, we analyze the joint effects of the

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


three variables, we note that they each predict different dimensions of the personal constructsystem. First, age predicts both the ways in which children construct their own representa-tional system and the specific contents of this system, especially with regard to the level ofabstraction and type of construct used. Gender, on the other hand, seems to have more limitedpredictive power than the other two variables, acting predominantly on the construing processand, in terms of contents, on difficulty in constructing a self-image and expression of emotion.Finally, attachment appears to mainly influence stability of self-image and representation ofrelationships, and in consequence, the emotional aspects of identity; this leads to specific diffi-culties on the part of insecurely attached children in construing a stable and positive constructsystem.

Finally, some brief methodological remarks regarding the limitations of this study and its possi-ble further developments. Although this study suggests that application of the self-characterizationtechnique with school-age children may be a valid means of exploring children’s internal world(at least with nonclinical subjects), it would be of value to conduct follow-up research to providea more exhaustive and systematic examination of the implications of using this method withchildren. The relatively limited size of our sample, especially considering the broad age rangeconsidered (from 7 to 11) and the homogeneous cultural background of the subjects (Italian), doesnot allow us to generalize from our findings. Furthermore, other factors may influence self/otherrepresentations, such as socialization models (Shaffer, 1984; Ugazio & Castiglioni, 1998) andmultiple attachments (van Ijzendoorn, Sagi, & Lambermon, 1992). In addition, our study did notexamine disorganized attachment style in children, a topic on which we are already working usingself-characterization with children displaying this attachment who have experienced traumaticevents (Veronese, Castiglioni, & Said, 2010).

Given that self-characterization was born in and specifically developed for clinical settings,we recommend its application to clinical samples, both for constructivist diagnosis and forplanning clinical intervention: Writing a self-characterization may help children to become moreaware of, and to reflect on, core aspects of their perceptions of self and others, as well as theirconcerns and fears, with key benefits for their psychophysical wellbeing (Campbell & Pennebaker,2003), including in the context of family therapy sessions involving children (Androutsopoulou,2001).

Given that self-characterization is a “quintessentially narrative technique” (Neimeyer, 1994, p.239), cross-comparison between it and other narrative techniques appears to be a promising lineof enquiry that could yield interesting data, also in relation to children. In particular, it could beof great interest to investigate how children’s self-characterizations correlate with their drawings(Bell & Bell, 2008; Procaccia, Veronese, & Castiglioni, in press), a line of research we intend topursue in the near future.

NOTES

1. This study was carried out in line with the Ethics Committee Guidelines of the University of Milano-Bicocca and approved by the Ethics Committee of MIUR (Italian Ministry of Education, University andResearch). Written parental informed consent was sought. The children themselves were also free to takepart in the research or to withdraw from it; similarly, they could decline to answer any of the questionsthey were asked.

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


2. Following presentation and description of each picture, children are asked four questions about a hypo-thetical child: How do you think the child in the picture feels? Why do you think he/she feels like this?What do you think the child will do now? What will the child do when he/she sees his/her mother (orparents) again?

3. In PCP, superordinate and comprehensive constructs are those with a higher level of generality—thatis to say, they are at a higher hierarchical level within the organization of the construct system andare therefore key to understanding the core meanings underpinning the entire structure (Kelly, 1955;Armezzani, 2010; Butt, 2008; Neimeyer, 2009).

4. In line with Art.18 on Privacy, Italian Law No.196/2003, “Code for the protection of personal informa-tion,” and subsequent amendments, all information identifying the authors of the self-characterizationshas been changed (names, places, etc.).

REFERENCES

Ainsworth, M., Blehar, M., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the strangesituation. Hillsdale, NJ: Erlbaum.

Androutsopoulou, A. (2001). The self-characterization as a narrative tool: Applications in therapy with individuals andfamilies. Family Process, 40, 79–94.

Armezzani, M. (2010). L’autocaratterizzazione: Uno strumento per la voce dell’altro [Self-characterization: A tool forthe other’s voice]. In M. Castiglioni & E. Faccio (Eds.), Costruttivismi in psicologia clinica (pp. 148–170). Torino:Utet.

Attili, G. (2001). Ansia da separazione e misura dell’attaccamento normale e patologico [Separation anxiety andmeasurement of normal and pathological attachment]. Milano: Unicopli.

Bannister, D., & Fransella, F. (1980). Inquiring man. Harmondsworth, UK: Penguin.Bell, S. J., & Bell, R. C. (2008). An illustration of self characterization in a child’s drawing: The importance of process.

Personal Construct Theory & Practice, 5, 1–9.Bieri, J., Bradburn, W. M., & Galinsky, M. D. (1958). Sex differences in perceptual behavior. Journal of Personality, 26,

1–12.Bowlby, J. (1969). Attachment and loss, vol. 1. Attachment. New York: Basic Books.Bowlby, J. (1973). Attachment and loss, vol. 2. Separation: Anxiety and anger. New York: Basic Books.Bretherton, I., & Munholland, K. A. (2008). Internal working models in attachment relationships: Elaborating a central

construct in attachment theory. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment: Theory, research, andclinical applications (2nd ed.; pp. 102–127). New York: Guilford.

Butt, T. (2008). George Kelly. Houndmills, UK: Palgrave Macmillan.Campbell, R. S., & Pennebaker, J. W. (2003). The secret life of pronouns: Flexibility in writing style and physical health.

Psychological Science, 14, 60–65.Cheshire, J. (2000). The telling or the tale? Narratives and gender in adolescent friendship networks. Journal of Sociolin-

guistic, 4, 234–262.Crittenden, N., & Ashkar, C. (2011). The self-characterization technique: Uses, analysis, and elaboration. In P. Caputi,

L. L. Viney, B. M. Walker, & N. Crittenden (Eds.), Personal construct methodology (pp. 109–128). Chichester, UK:Wiley-Blackwell.

Denicolo, P. (2003). Elicitation methods to fit different purposes. In F. Fransella (Ed.), International handbook of personalconstruct psychology (pp. 123–131). Chichester, UK: Wiley.

Feixas, G., & Villegas, M. (1991). Personal construct analysis of autobiographical texts. International Journal of PersonalConstructs Psychology, 4, 51–83.

Fivush, R. (2011). The development of autobiographical memory. Annual Review of Psychology, 62, 559–582.Fivush, R., Bohanek, J., Zaman, W., & Graphin, S. (2012). Gender differences in adolescents’ autobiographical narratives.

Journal of Cognition and Development, 13, 295–319.Fivush, R., & Haden, C. A. (2003). Autobiographical memory and the construction of a narrative Self: Developmental

and cultural perspectives. Mahwah, NJ: Lawrence Erlbaum.

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


Fonagy, P., & Target, M. (1997). Attachment and reflective function: Their role in self organization. Development andPsychopathology, 9, 679–700.

Fransella, F. (1995). George Kelly. London: Sage.Fransella, F., Bell, R. C., & Bannister, D. (2004). A manual for repertory grid technique (2nd ed.). Chichester, UK: Wiley.Fransella, F., & Dalton, P. (1990). Personal construct counselling in action. London: Sage.Green, B. (2004). Personal construct psychology and content analysis. Personal Construct Theory & Practice, 1, 81–92.Guidano, V. F. (1987). Complexity of the self. New York: Guilford.Harris, P. (1989). Children and emotions. Oxford: Blackwell.Hayhow, R., Lansdown, R., Maddick, J., & Ravenette T. (1988). PCP and children. In F. Fransella & L. Thomas (Eds.),

Experimenting with personal construct psychology (pp. 199–209). London: Routledge.Jackson, S. R. (1988). Self-characterization: Dimensions of meaning. In F. Fransella & L. Thomas (Eds.), Experimenting

with personal construct psychology (pp. 223–231). London: Routledge.Jackson, S. R. (1990). Self-characterization: Development and deviance in adolescent construing. In P. Maitland (Ed.),

Personal construct theory, deviancy, and social work (pp. 60–68). London: Inner London Probation Service & Centrefor Personal Construct Psychology.

Jankowicz, D. (2004). The easy guide to repertory grids. Chichester, UK: Wiley.Kelly, G. A. (1955). The psychology of personal construct. New York: Norton.Klagsbrun, M., & Bowlby, J. (1976). Responses to separation from parents: A clinical test for young children. Projective

Psychology, 21, 109–126.Landfield, A. W. (1971). Personal construct systems in psychotherapy. Chicago: Rand McNally.Leung, K., Marsh, H., Craven, R., Yeung, A., & Abduljabbar, A. (2013). Domain specificity between peer support and

self-concept. Journal of Early Adolescence, 33, 227–244.Liotti, G. (1991). Patterns of attachment and the assessment of interpersonal schemata. Journal of Cognitive Psychother-

apy, 5, 105–115.Little, B. R. (1968). Factors affecting the use of psychological vs. non-psychological constructs on the rep test. Bulletin

of the British Psychological Society, 21, 34.Mancuso, J. (2003). Children’s development of personal construct. In F. Fransella (Ed.), International handbook of

personal construct psychology (pp. 275–282). Chichester, UK: Wiley.Main, M. (1991). Metacognitive knowledge, metacognitive monitoring, and singular (coherent) vs/ multiple (incoherent)

models of attachment. In C. M. Parkes, J. Stevenson-Hinde, & P. Marris (Eds.), Attachment across the life cycle (pp.127–159). London & New York: Routledge.

Neimeyer, R. A. (1994). The role of client-generated narratives in psychotherapy. Journal of Constructivist Psychology,7, 229–242.

Neimeyer, R. A. (2009). Constructivist psychotherapy. London & New York: Routledge.Neimeyer, R. A., & Mahoney, M. J. (Eds.). (1995). Constructivism in psychotherapy. Washington, DC: American

Psychological Association.Nelson, K., & Fivush, R. (2004). The emergence of autobiographical memory: A social cultural developmental theory.

Psychological Review, 111, 486–511.Parkes, C. M., Stevenson-Hinde, J., & Marris, P. (Eds.). (1991). Attachment across the life cycle. London & New York:

Routledge.Procaccia, R., & Castiglioni, M. (2010). La costruzione dell’immagine di Se e degli altri in eta evolutiva: Uno studio

pilota attraverso l’autocaratterizzazione [The construction of self-other representation in children: A pilot study usingself-characterization]. In M. Castiglioni & E. Faccio (Eds.), Costruttivismi in psicologia clinica (pp. 353–377). Torino:Utet.

Procaccia, R., & Di Blasio, P. (2009). Narrazioni post-traumatiche nei bambini maltrattati e abusati [Post-traumaticnarrations in mistreated and abused children]. In V. Ardino (Ed.), Il disturbo post-traumatico da stress nello sviluppo(pp. 19–42). Milano: Unicopli.

Procaccia, R., Veronese, G., & Castiglioni, M. (In press). The impact of attachment style on the family drawings ofschool-aged children. Open Psychology Journal.

Procter, H. (1981). Family construct psychology: An approach to understanding and treating families. In S. Walrond-Skinner (Ed.), Development in family therapy (pp. 350–366). London: Routledge.

Ravenette, T. (1977). Personal construct theory: An approach to the psychological investigation of children and youngpeople. In A. T. Ravenette (Ed.), Tom Ravenette: The selected papers. Personal construct psychology and the practiceof an educational psychologist (pp. 47–79). Farnborough, Hampshire: EPCA Publications.

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014


Ravenette, T. (2003). Constructive intervention when children are presented as problems. In F. Fransella (Ed.), Interna-tional handbook of personal construct psychology (pp. 283–293). Chichester, UK: Wiley.

Salmon, P. (1970). A psychology of personal growth. In D. Bannister (Ed.), Perspective in personal construct theory(pp. 197–222). London: Academic.

Shaffer, H. R. (1984). The child’s entry into a social world. London: Academic.Singer, J., Blagov, P., Berry, M., & Oost, M. (2013). Self defining memories, scripts, and the life story: Narrative identity

in personality and psychotherapy. Journal of Personality, doi: 10.1111/jopy.12005.Trevarthen, C., & Aitken, K. J. (2001). Infant intersubjectivity: Research, theory, and clinical application. Journal of

Child Psychology and Psychiatry, 42, 3–48.Ugazio, V. (2013). Semantic polarities and psychopathologies in the family: Permitted and forbidden stories. New York:

Routledge.Ugazio, V., & Castiglioni, M. (1998). Socialization models and the construction of self. Journal of Constructivist

Psychology, 11, 3–29.van Ijzendoorn, M. H., Sagi, A., & Lambermon, M. W. E. (1992). The multiple caretaker paradox: Data from Holland

and Israel. New Directions for Child and Adolescent Development, 57, 5–24.Veronese, G., Castiglioni, M., & Said, M. (2010). The use of narrative-experiential instruments in contexts of military

violence: The case of Palestinian children in the West Bank. Counselling Psychology Quarterly, 23, 411–423.Veronese, G., Procaccia, R., Romaioli, D., Barola, G., & Castiglioni, M. (2013). Psychopathological organizations and

attachment styles in patients suffering of fear of flying: A case study. Open Psychology Journal, 6, 20–27. doi:10.2174/1874350101306010020.

Veronese, G., Romaioli, D., & Castiglioni, M. (2012). Attachment styles and construction of self in a clinical group ofaerophobic: A pilot study. Psychological Studies, 57, 303–309. doi: 10.1007/s12646-011-0136-4.

Wallin, D. J. (2007). Attachment in psychotherapy. New York: Guilford.White, M., & Epston, D. (1990). Narrative means to therapeutic ends. New York: Norton.Zaman, W., & Fivush, R. (2013a). Gender differences in elaborative parent–child emotion and play narratives. Sex Roles,

6, 591–604.Zaman, W., & Fivush, R. (2013b). Stories of parents and self: Relations to adolescent attachment. Developmental

psychology, 49, 2047–2056. doi: 10.1037/a0032023.

Dow

nloa

ded

by [

Uni

vers

ita' M

ilano

Bic

occa

], [

Mar

co C

astig

lioni

] at

07:

36 1

6 M

ay 2

014

Self-Characterization and Attachment Style: A Creative Method of Investigating Children's Construing

Documents

Transcript of Self-Characterization and Attachment Style: A Creative Method of Investigating Children's Construing