The Expression of Attachment History

1

Kõiv, Kristi

The Expression of Attachment History

in the Kinetic Family Drawings of Antisocial Children

Paper presented at the European Conference on Educational Research, Edinburgh, 20–23 September 2000

A matched pair method was used to select antisocial children from all 312 the elementary school children in one school. A sample of 16 children with externalized antisocial behavior patterns was matched to equal number of unproblematic children by age, sex, school, class, marital status of parents, number of family members living together, residence location, economical status of family and parents' occupation. At first, the use of Kinetic Family Drawing variables as expressions of disturbances of different attachment history of antisocial children was subsequent: compartmentalism of figures; akinesis; omission of body parts; buttons; pencil erasures; heavy and overworked line quality; high activity level of father figure; self-figure closer to the father figure than to the mother figure. Secondly, the expression of attachment style beyond infancy and childhood was supported by this empirical research: subjects endorsing the secure description reported overall childhood care and autonomy from parents; avoidant and anxious-ambivalent subjects perceived disturbances in parental bonding as rejecting and overprotection in childhood.

Introduction

J. Bowlby’s (1969) theory of attachment deals primarily with the bonds formed between infants and their caregivers. The extension of attachment principles beyond infancy and childhood were supported by theoretical work and empirical studies carried out during last decades (Freeney & Noller, 1996; Sperling & Berman, 1994).

C. Hazan and P. R. Shaver’s (1987) initial papers provide an empirical support to the attachment perspective on romantic love consisting of two studies of adult samples investigating the association between attachment style and aspects of childhood and adult relationships. They found that adults with different attachment styles experience their most important love relationships differently and recall their childhood relationships with parents differently.

Retrospective accounts of attachment experiences obtained by self-reports allow examining the dimensions of attachment on the basis of normative and pathological attachment processes of adults and develop instrument for assessing individuals’ retrospective experience of their parental attachment relationships (Parental Bonding Instrument; Parker et al., 1979). These factor-analytically derived dimensions of parental care and protection define the parental contribution to bonding as reflecting key dimensions underlying parental attitudes and behaviors.

Parental unresponsiveness has been conceptualized by attachment theorists as being most critical to the development of self-regulation skills of children. Differences in caregiver’s sensitivity and the resultant bond quality between the parent and a child is an important factor in developing later patterns of child behaviour (Bowlby, 1969). The view that children are predisposed to form attachment during infancy has considerable importance for the study of development of child problems. The literature contains numerous reports of samples of high-risk families indicated that insecure attachment during infancy was associated with an increating risk of later externalizing problems of children (Shaw & Vondra, 1995; Erickson et al., 1985; Lyons-Ruth et al., 1993; Renken et al., 1989; Troy & Sroufe, 1987).

Although, during the past three decades, Kinetic Family Drawings (KFD; Burns & Kaufman, 1970, 1972) have been used to recognize the important role of family dynamics in the etiology and treatment of emotional disorders of children (see reviews: Cummings, 1986; Handler & Habenicht, 1994, Knoff & Prout, 1985), there have been done a few studies in the field of associations of children’s externalizing antisocial behaviour and its expression in the family projective drawings. For example manifestations of antisocial behaviour in legal perspective had been the subject of an empirical investigation about adolescent delinquents KFDs: Sobel & Sobel (1976) compared adolescent delinquents with a control group and found that delinquents did not use kinesis between family members, omitted family members and bodies of people. On the other hand, manifestations of antisocial perspective in psychopathological perspective had been a subject of O’Brien and Patton’s (1974) comparative study of fourth- through eighth-grade children with scores on behavioral checklist completed out by teachers, and anxiety, self-esteem inventories for children.

Indeed, Fury et al. (1997) had suggested that middle childhood high-risk children’s family drawings were effective measures of representational models of attachment, there were no reports about connections between antisocial children’s attachment history and its expression in the projective family drawings of children.

The aim of the present study was to find a link between antisocial children’s attachment history and its expressions in the Kinetic Family Drawings.

Based on previous research it was predicted that disturbances of bonding of antisocial children with parents would manifest on children’s KFDs as total absence of family members, omission of body parts and akinesis.

Method

Selection of subjects

Rating scales about different kinds of antisocial behavior based on Frick et al. (1993) meta-analytic research were used to determine the various externalized behavior patterns of antisocial behavior. Teachers filled in rating scales, which contained items about four subclasses of externalized actions of antisocial behavior: property violation (cruel to animals, firesetting, lies, steals, vandalism), status violation (breaks rules, runaway, substance abuse, swears, truancy), oppositional (annoys, argues, angry, defies, stubborn, temper, touchy), aggression (assault, blames others, bullies, cruel, fights, spiteful). Rating scales required teachers to judge a child’s behavioural patterns in terms of severity: does not appear, sometimes, or frequently appears. Only frequently occurred items were scored to identify “typical” antisocial children - the most frequent pattern of antisocial behavior of children selected from ratings of all children. “Typical” antisocial children (antisocial children) had a variety of externalized problems: argues, breaks rules, bullies, cruel, defies, fights, lies, steals, stubborn, substance abuse, swears, temper, truancy.

A sample of 16 antisocial children was selected from all 312 elementary school children and matched with control group children. Antisocial children were matched by age, sex, school, class, marital status of parents, number of family members living together, residence location, economical status of family and parents' occupation with control group.

Subjects of this study were antisocial children and their matched controls and their parents. A total number of children from different group was 32 children (28 boys and 4 girls) ranging from 8-10 yeas in age (M=9,9; SD=5,74) and a total number of parents from two groups was 60 (32 mothers and 28 fathers) ranging from 25-54 years old (M=37,2; SD=10,14) (four families were incomplete).

Instruments

Multiple-item attachment scale, developed by Simpson (1990), was used to define in terms of present reports of attachment. This measure based directly on Hazan and Shaver (1987) attachment measure indicating three paragraphs corresponding to the three attachment styles: secure, avoidant, anxious-ambivalent. Antisocial and matched children and their parents were asked to rate 13 sentences contained within the Hazan and Shaver (1987) adult attachment measure on a 7-point Likert-type scale, ranging from strongly disagree (1) to strongly agree (7).To measure each attachment style, the items corresponding to three paragraphs aggregated to form three attachment indexes.

Parental Bonding Instrument (PBI; Parker et al., 1979) was used for measuring two different parental bonding dimensions - care and overprotection including orthogonal constructs: care-rejection and overprotection-autonomy.

The four quadrants were created by the intersection of these two bipolar dimensions: high autonomy and care is called optimal parenting; high overprotection and care is termed affectionate constraint; high overprotection and rejection is called affectionless control; high autonomy and rejection is named as absent or weak bonding. Antisocial and matched control group children and their parents were asked to rate 25-item PBI questionnaire measuring the parental characteristics of care and overprotection.

Kinetic Family Drawing (KFD). The KFDs were administered in accordance with the instructions described by Burns and Kaufman (1972): “Draw a picture of everyone in your family, including yourself, doing something. Try to draw whole people, not just cartoons or stick people. Remember, make everyone doing something – some kind of action.”

The KFD drawings of antisocial and matched children were scored using two quantitative methods. At first, Reynold’s guide (1978) was adapted into the scoring system, consisting of 37 individual indicators that are scored for presence or absence. Secondly, O’Brien and Patton (1974) scoring system using 29 individual variables (inter-figure distance; figure size; barriers between figures; activity level and orientation of major figures) were employed to score KFDs obtained from two different groups of children.

Antisocial and matched children and their parents filled in a questionnaire describing on a 4-point Likert-type scale the degree of satisfaction they attributed to the present achievement, relations with close individuals and him/herself ranging from very satisfying (1) to very dissatisfying(4). One additional feature – long-time separation from parents was introduced into thequestionnaire.

Results

Factor analysis was applied to examine the internal structure of different attachment styles and parental bonding dimensions of antisocial and matched children and their parents and to see connections with variables of KFDs. The Varimax rotated matrices of factor loadins obtained for each of the four used research methods variables of antisocial and matched controls. The loadings of variables on 7 factors, which account for a cumulative total of 54,2% of variance, are presented in Table 1. The seven factors accounted for 12,9%, 8,7%, 7,7%, 7,1%, 6,7%, 5,8% and 5,3% of the total variation.

VARIABLES

/

F a c t o r l o a d i n g s

1
factor / 2
factor / 3
factor / 4
factor / 5
factor / 6
factor / 7 factor
secure attachment of mother / 0,541 / 0,354 / -0,170 / -0,063 / -0,292 / -0,009 / 0,109
secure attachment of father / 0,582 / 0,200 / -0,002 / -0,121 / -0,294 / -0,040 / 0,277
maternal care / 0,649 / 0,818 / -0,818 / -0,112 / -0,261 / -0,177 / 0,269
maternal autonomy / 0,624 / 0,259 / -0,001 / -0,231 / -0,212 / 0,003 / 0,107
paternal care / 0,634 / 0,002 / 0,001 / -0,265 / 0,003 / -0,178 / 0.133
paternal autonomy / 0,546 / 0,026 / 0,002 / -0,187 / 0,002 / 0,009 / 0,110
maternal autonomy of mother in childhood / 0,446 / 0,164 / -0,164 / -0,308 / -0,202 / -0,159 / 0,383
maternal autonomy of father in childhood / 0,467 / 0,128 / -0,251 / 0,006 / 0,008 / -0,214 / 0,127
maternal care of mother in childhood / 0,432 / 0,120 / 0,013 / 0,171 / 0,101 / 0,033 / 0,004
satisfaction with relationship of child / 0,508 / -0,020 / -0,004 / 0,083 / -0,164 / 0,001 / 0,233
satisfaction with achievement of father / 0,292 / 0,572 / -0,194 / -0,119 / -0,147 / -0,247 / 0,123
separation of father from parents in childhood / -0,008 / -0,661 / 0,023 / 0,132 / 0,143 / 0,013 / -0,004
paternal care of mother in childhood / 0,361 / 0,792 / 0,006 / 0,008 / 0,116 / -0,166 / 0,003
paternal care of father in childhood / 0,343 / 0,485 / 0,001 / -0,006 / -0,002 / -0,165 / 0,008
paternal autonomy of mother in childhood / 0,392 / 0,517 / -0,104 / -0,192 / -0,005 / -0,101 / 0,220
paternal autonomy of father in childhood / 0,367 / 0,553 / -0,144 / -0,122 / 0,001 / 0,001 / 0,009
satisfaction with him/herself of child / 0,105 / 0,398 / -0,116 / -0,090 / -0,285 / -0,135 / 0,248
maternal overprotection / 0,005 / -0,004 / 0,498 / 0,009 / 0,003 / 0,198 / -0,311
paternal overprotection / 0,009 / 0,007 / 0,486 / 0,007 / 0,009 / -0,001 / -0,004
maternal rejection / -0,005 / 0,105 / 0,418 / 0,162 / 0,005 / -0,136 / -0,247
paternal rejection / -0,009 / -0,207 / 0,524 / -0,006 / 0,012 / -0,234 / -0,346
maternal care of father in childhood / 0,452 / 0,017 / -0,711 / -0,201 / 0,005 / 0,002 / -0,031
avoidant attachment of father / -0,003 / -0,398 / 0,138 / 0,472 / 0,128 / 0,202 / -0,041
paternal overprotection of mother in childhood / -0,257 / 0,082 / 0,254 / 0,585 / 0,006 / 0,114 / -0,002
paternal rejection of mother in childhood / -0,078 / -0,238 / -0,007 / 0,497 / 0,007 / 0,144 / -0,176
anxious-ambivalent attachment of father / -0,091 / -0,002 / 0,006 / 0,007 / 0,451 / 0,233 / -0,006
anxious-ambivalent attachment of mother / -0,008 / -0,001 / 0,007 / 0,061 / 0,523 / 0,256 / -0,121
separation of mother from parents in childhood / -0,008 / -0,106 / 0,245 / -0,205 / 0,378 / 0,122 / 0,009
paternal overprotection of father in childhood / -0,257 / 0,008 / 0,254 / 0,241 / 0,603 / 0,301 / 0,002
paternal rejection of father in childhood / -0,245 / -0,100 / 0,132 / 0,008 / 0,428 / 0,184 / -0,003
maternal overprotection of mother in childhood / -0,142 / -0,123 / 0,112 / 0,176 / 0,655 / 0,001 / -0,204
maternal rejection of mother in childhood / 0,004 / -0,124 / 0,248 / 0,011 / 0,455 / 0,227 / -0,209
avoidant attachment of mother / -0,147 / -0,173 / 0,149 / 0,131 / 0,323 / 0,627 / -0,102
maternal overprotection of father in childhood / -0,161 / -0,138 / 0,107 / 0,003 / 0,295 / 0,442 / -0,331
maternal rejection of father in childhood / -0,175 / -0,300 / 0,298 / 0,001 / 0,175 / 0,495 / -0,324
satisfaction with achievements of child / 0,323 / 0,117 / -0,272 / -0,123 / -0,001 / -0,413 / -0,009
satisfaction with relationships of mother / 0,177 / 0,221 / 0,007 / -0,002 / -0,007 / -0,002 / 0,460
satisfaction with relationships of father / 0,260 / 0,332 / 0,001 / 0,001 / -0,008 / -0,003 / 0,495
satisfaction with herself of mother / 0,146 / -0,006 / -0,008 / 0,121 / -0,189 / -0,008 / 0,445
satisfaction with himself of father / 0,135 / 0,161 / -0,246 / -0,120 / -0,291 / -0,291 / 0,480
satisfaction with achievement of mother / 0,126 / 0,006 / -0,227 / -0,005 / -0,247 / -0,126 / 0,381
separation of child from parents / -0,010 / -0,183 / 0,225 / 0,180 / 0,192 / 0,007 / -0,681
self-figure closer to the father figure than to the mother / -0,183 / -0,266 / 0,566 / 0,004 / 0,129 / 0,230 / 0,006
heavy and overworked line / 0,232 / 0,268 / -0,102 / -0,206 / -0,242 / -0,209 / -0,466
high activity level of father / 0,243 / 0,230 / -0,207 / -0,005 / -0,003 / -0,286 / -0,450
pencil erasures / -0,211 / -0,194 / 0,261 / 0,404 / 0,296 / 0,269 / -0,001
compartmentalism of figures / -0,206 / -0,300 / 0,006 / 0,111 / 0,367 / 0,201 / -0,286
akinesis of figures / -0,009 / -0,007 / 0,115 / 0,112 / 0,398 / 0,006 / -0,274
omission of body parts / -0,146 / -0,119 / 0,255 / 0,003 / 0,486 / 0,239 / 0,008
buttons / -0,211 / -0,193 / 0,007 / 0,005 / 0,425 / 0,009 / 0,003

Table 1. Factor loadings for 50 variables (multiple-item attachment scale, Parental Bonding Instrument, questionnaire, Kinetic Family Drawing (only high loadings were included)).

The first factor might be termed as secure attachment of parents. It was defined by the following variables: secure attachment of mother and father; optimal bonding (high care and autonomy) of parents; parental reports of childhood optimal bonding; child’s satisfaction with close relationships.

Factor II seems to be a factor of afathers’ satisfaction with his achievement and it was linked to the fathers’ satisfaction with his achievement, paternal reports of childhood optimal bonding of parents, less separation of father from his parents in childhood and child’s satisfaction with him/herself.

Factor III might be termed as affectionless control of parents. It is defined by the parents’ affectionless control (overprotection and rejection), low maternal reports of childhood care from the parent of opposite sex and variable in the KFDs that child drew the self-figure closer to the father figure than to the mother figure.

Factor IV is a factor of avoidant attachment of father. This name was intended as a convenient label because highest loadings were subsequent: avoidant attachment of the father; paternal reports of childhood affectionless control from the parent of opposite sex; variable in KFDs: pencil erasures.

Factor V named anxious-ambivalent attachment of parents. It is characterized by the anxious-ambivalent attachment of parents, maternal and paternal reports of childhood affectionless control from the parent of same sex, separation of mother from her parents in childhood and KFDs variables: compartmentalism of figures (one or more straight lines used to separate one or more figures), akinesis (motionless figures), missing essential body parts, buttons (overemphasized).

Factor VI was named avoidant attachment of mother because three highest loadings were: the avoidant attachment of mother; maternal reports of childhood affectionless control from the parent of opposite sex; child dissatisfaction with his/her achievement.

The following variables characterized Factor VII: the satisfaction of father and mother with him/herself and close relationships; less child long-time separation from parents; KFDs variables: less heavy and overworked line quality, low activity level of father figure. The last factor was labelled parents’ satisfaction with close relationships and themselves.

To test the differences among the subjects groups, Fisher F-criterion or Chi Square Tests were computed for each variable. Table 2 presents comparisons between antisocial and matched control children and their parents.

V A R I A B L E S / Antisocial children /

Matched

controls
Mean / Mean / F / p <
Parental overprotection (PBI) / 14,9 / 3,8 / 3,9 / 0,05
Maternal care (PBI) / 8,5 / 20,6 / 3,9 / 0,05
Avoidant attachment of mother (attachment scale) / 2,6 / 21,4 / 3,8 / 0,05
Avoidant attachment of father (attachment scale) / 3,1 / 24,4 / 3,6 / 0,05
Anxious-ambivalent attachment of mother (attachment scale) / 5,2 / 24,7 / 6,7 / 0,05
Anxious-ambivalent attachment of father (attachment scale) / 6,9 / 21,7 / 5,2 / 0,05
Activity level of father (KFD) / 6,2 / 1,1 / 8,6 / 0,001
The distance from self to father (KFD) / 2,03 / 9,2 / 5,0 / 0,05
V A R I A B L E S / Antisocial children /

Matched

controls
% / % / 2 / p <
Separation of child from parents in childhood (questionnaire) / 81,3 / 32,3 / 8,13 / 0,05
Pencil erasures of father figure (KFD) / 100 / 18,8 / 21,9 / 0,001
Missing essential body parts (KFD) / 100 / 0 / 32,0 / 0,001
Compartmentalism of figures (KFD) / 93,8 / 6,3 / 24,5 / 0,001
Line quality heavy and overworked(KFD) / 87,5 / 0 / 24,9 / 0,001
Motionless or stick figures (akinesis) (KFD) / 87,5 / 0 / 24.9 / 0,001
Buttons (KFD) / 87,5 / 0 / 24,9 / 0,001

Table 2. Differences between antisocial and control groups measured by multiple-item attachment scale, PBI, KFD and questionnaire.

Two statistically significant Parental Bonding Instrument scale scores distinguishing the groups were parental overprotection and maternal care, with the antisocial children reporting their parents as more overprotective and mother as less caring than matched controls. Antisocial children and their mothers had a higher incidence of a long-term separation from parents in childhood than control groups. There were significant differences in avoidant and anxious-ambivalent attachment of mother and father between the two groups. When compared to the matched control population, KFDs done by antisocial children had more examples of pencil erasure of father figure, omission of body parts, compartmentalism of figures, akinesis, omission of body parts, buttons, small distance between self-figure and father figure versus high distance, heavy and overworked line versus no heavy line and high activity level of the father figure versus low activity level.

Discussion

At first, the results indicated that parental secure attachment was connected with parental optimal bonding (high ratings in care and autonomy) and maternal and paternal reports of optimal bonding in childhood.

High ratings in care dimension had been interpreted as affection, emotional warmth, empathy, closeness and high ratings in autonomy as allowance of independence and autonomy (Parker et al., 1979).

Early studies of adult attachment history showed that secure individuals tend to remember their parents as warm and affectionate (Collins & Read, 1990; Freeney & Noller, 1990; Hazan & Shaver, 1987).

Persent results asserted and specified previous research: secure adult attachment style had connections not only with reports of childhood care, but also with autonomy from parents. The early parental optimal bonding had an effect to the current secure attachment relationships of parents influence on child’s satisfaction with close relationships.

The importance of optimal bonding of parents was especially significant for fathers -

paternal reports of childhood optimal bonding and a fact that father had less experiences of separating from his parents in childhood influenced in adulthood males fulfillment of achievements. The satisfaction of achievement of father affected his offspring’s satisfaction with him/herself. This finding is connected with the proposition that beliefs and attitudes of secure subjects are consistent with their higher self-esteem and self-confidence (Freeny & Noller, 1990).

Consequently, several qualities of early attachment history mediated the current attachment relationships and paternal satisfaction with achievement and child’s satisfaction with relations with close individuals and with him/herself.

Secondly, the parental affectionless control (high ratings in overprotection and rejection) as bonding style and low maternal reports of childhood care (rejection) from the parent of opposite sex were connected and had expression in the KFDs of children as a characteristic that child drew the self-figure closer to the father figure than to the mother figure.