You are on page 1of 8

Jolly R., Vulić-Prtorić A. – XI. DANI PSIHOLOGIJE U ZADRU, 27.-29. svibnja 1998.

RICHARD JOLLEY
Staffordshire University, UK
&
ANITA VULIĆ-PRTORIĆ
Odsjek za psihologiju, Filozofski fakultet, Zadar

SIZE AND PLACEMENT IN CHILDREN'S DRAWINGS AS


CORRELATES OF EMOTIONAL DISTRESS
XI. DANI PSIHOLOGIJE U ZADRU

Sixty Croatian children who had experienced directly the recent war in Croatia drew a
man, followed on a separate page by either (i) a Croatian soldier, (ii) an enemy soldier or (ii)
a second drawing of a man. Half of the sample had lost their father due to the war. There were
no significant effects of Trauma group (with/ without father) or Topic of drawing on the
drawings' size or their placement relative to the child's self-portrait drawing. The paper
discusses the many mechanisms of influence of emotion on drawing but argues that these are
idiosyncratic and unpredictable.

The use of drawings in clinical assessments of adults and children has had a long
tradition. Indeed, diagnostic drawing tests still rank in the top ten of all tests used by
clinicians (Lubin, Larsen, Matarazzo & Sever, 1985; Watkins, Campbell, Nieberding &
Hallmark, 1995). Furthermore, there are numerous case studies using drawings as a
therapeutic tool (e.g., see Dalley, 1984; Winner, 1982). Both forms of practice indicate that it
is commonly considered by those working with disturbed clients that drawings can provide a
useful insight into the client's problems. Although there is little dispute that even line
drawings can be expressive using aesthetic criteria (Arnheim, 1949; Gardner, 1974; Werner &
Kaplan, 1963), the evidence for drawings conveying an emotional maladjustment in the
drawer is far less convincing.

Theoretical links between drawings and the drawer's maladjustment can be broadly
categorized under three traditions (see Cox, 1992; Thomas & Silk, 1990; Thomas & Jolley, in
press). First, Machover (1949) claimed that drawings can be used as a personality assessment.
She devised the "Draw-a-Person" test in which the client is asked to draw a person and then
make a second drawing of a person of the opposite gender to that depicted in the first
drawing.
Interpretations of the drawings are then made by the clinician using psychoanalytic theory. In
extensive reviews of the literature, Swenson (1968) and Roback (1968) argue that the Draw-
a-Person Test has weak reliability and validity, and that particular aspects of a drawing
cannot be related to particular types of maladjusted personality (see also Falk, 1981). This is
likely to be due to the tradition's reliance on the body-image assumption. That is, a drawing of
an unidentified person conveys the artist's self-concept and possibly their physical image.
Jolly R., Vulić-Prtorić A. – XI. DANI PSIHOLOGIJE U ZADRU, 27.-29. svibnja 1998. 2

This validity of this assumption has yet to be established (Golomb, 1992; Mortenson, 1991;
Roback, 1968; Swensen, 1968; Thomas & Jolley, in press).

An alternative approach was adopted by Koppitz (1968, 1984) who claimed that
features in children's human figure drawings convey the children's current emotional attitudes
not only about themselves but also towards the person depicted. The assertion that drawings
can also convey the child's feelings towards the topic places less reliance on the body-image
hypothesis. In addition, drawing is more of a natural and uninhibited experience for children
and therefore children's drawings may be a more valid diagnostic tool than using adult
drawings. In the Koppitz test the child is asked to draw a person, unspecified by the
interviewer. Koppitz produced 30 emotional indicators that she argued were atypical in the
drawings of normal children but were consistent features of emotionally disturbed
children. Although a few studies have reported findings supporting the validity of the
emotional indicators (e.g., Sturner, Rothbaum, Visitainer & Wolfer, 1980), others have
reported overall scores that have not differentiated between emotionally disturbed and control
groups (e.g., Eno, Elliot & Woehlke, 1981; Forrest & Thomas, 1991; Fuller, Preuss &
Hawkins, 1970). Although Koppitz (1968) questioned the validity of the body-image
hypothesis her methodology of obtaining a drawing of an unspecified person to comment
upon the child's current emotional state does not sufficiently distance itself from the
hypothesis.

In contrast, children's drawings often depict the important people in their lives and it is
not unreasonable to conceive that these familiar topics convey the child's attitude towards
these topics. Applied to a clinical setting, the child may be asked to draw specific people
known to the child, as in the "Kinetic Family Drawing Test" (Burns & Kaufman, 1970). The
basis for such tests is derived from Lowenfeld's (1939) claim that children exaggerate features
of a topic that are important to them (see also Lowenfed & Brittain, 1975). This claim was
supported in a series of studies of Santa Claus drawings (e.g., Craddick 1963; Sechrest &
Wallace, 1972) in which children's drawings of Santa Claus tended to be larger as Christmas
approached and then reduced in size after the Christmas period. The importance hypothesis
has been extended in more recent studies to show that aversive topics are drawn significantly
smaller than neutral topics, possibly as a result of a defence mechanism (see Fox & Thomas,
1990; Thomas, Chaigne & Fox, 1989. Using a more methodologically sophisticated design
than reported in the earlier studies, Thomas, Chaigne & Fox (1989) asked children to copy a
shape of a man, and then make a second drawing of the man characterized as either nice,
nasty or neutral. The attempt in this study, therefore, was to manipulate the importance of the
characters through the instructions. Also, children were encouraged to draw only the shapes
of the characters without including details as the inclusion of details may in itself increase the
size of the depiction (Freeman, 1980; Henderson & Thomas, 1990).

The adaptations made in this more recent work, however, are not without their own
problems. The instruction to omit any details could have appeared strange to children of the
age group tested (5- to 7-year-olds) who often include details in their spontaneous drawings.
In addition, providing character descriptions may not provide a valid test of the importance
hypothesis. Merely describing a character is unlikely to make that character meaningful to the
child. Thus, the topics drawn may not have been important to the child in the same way as the
real people in their lives. Note that this criticism is similar to the one made earlier in respect
of the body-image hypothesis. If children do emphasize or reduce features of friendly or
Jolly R., Vulić-Prtorić A. – XI. DANI PSIHOLOGIJE U ZADRU, 27.-29. svibnja 1998. 3

aversive topics that are important to them then surely this will occur most reliably for topics
of a more personal nature to the child. Indeed, in a series of seven studies that also
manipulated character descriptions through the instructions, Jolley (1995) was unable to
replicate the relationship between size and topic characterization found in the Thomas'
studies.

A personal link between the drawn topic and child was adopted by Thomas & Gray
(1992) who asked children to draw someone they liked and disliked next to a drawing of
themselves. Placement of figure has also been a commonly cited feature by which children
expressive their feelings (Koppitz, 1968; Machover, 1949). Thomas & Gray reported that
although liked figures were drawn closer to the self-potrait than disliked figures, the effect
disappeared in some drawing sequences. It is likely that the effect of emotional attitude on
size and placement of topic is stronger and more reliable in cases where the child has strong
feelings towards that topic, such as in instances of emotional distress.

This issue was partly addressed by Joiner, Schmidt & Barnett (1996) who obtained a
variety of depression and anxiety measures from 80 child and adolescent psychiatric
inpatients. Although the authors did not state they were testing the importance hypothesis, at
least two of the drawings tests administered had been derived from this tradition (e.g.,
"Kinetic Family Drawing Test" and the "Draw-a-Family" test).

The authors reported that the particular graphic features under investigation (size,
detail and line heaviness) were unrelated to the depression and anxiety measures. Different
findings may have been reported, however, if these children had been asked to draw the
source of their anxiety. The Joiner et al. paper implies that such information was not available
but in any event the sources of distress were likely to have been disparate.

One method of obtaining a large group of children who have all suffered from the
same source is to test children who have experienced war. Children are undoubtedly affected
negatively by war and its consequences. This is no more true of the recent war in Croatia in
which populations of the waring factions were living together in multi-national communities.
It was considered, therefore, that children in this country would continue to hold particularly
negative emotions about war. This is likely to be especially true for children who had lost
their father during the war.

Although, there are many studies that have reported war-based drawings,
interpretations have often been of a descriptive nature (e.g., Golomb, 1992). The present
study systematically investigated whether Croatian 7- to 10-year-olds differentiated their
drawings of emotive topics in terms of size and placement. All the children had lived in Zadar
whilst the town had been attacked from land, sea and air. Half the children had lost their
father as a result of the war. Due the length and extensiveness of the war it was not feasible to
obtain a control group of Croatian children that could be classified as being completely
unaffected emotionally by the war. The main topic of the drawings under investigation was
that of soldiers. Although the actions and subsequent emotional trauma of the war could
conceivably be assigned to the political leaders in the government from an adult's perspective,
children of the ages tested are more likely to attribute it more directly, that is, to the enemy
soldiers. Similarly, the Croatian soldiers are likely to be conceived as representing a
Jolly R., Vulić-Prtorić A. – XI. DANI PSIHOLOGIJE U ZADRU, 27.-29. svibnja 1998. 4

favourable personal importance to the Croatian children. All children were asked to draw a
man, followed by either another man (control condition), a Croatian (friendly) soldier or an
enemy soldier. Asking a sample of the children to draw a non-emotive topic provides an
important control comparison to the drawings of the emotive topics. Each drawing was
produced on a separate page so the amount of available space for each drawing was controlled
for. To emphasize the personal importance between the child and each topic drawn, a self-
portrait drawing was produced by the child and attached to each of the two pages in which the
men or soldiers were drawn. According to the importance hypothesis it was predicted that
enemy soldiers would be drawn smaller and further away from the self-portrait drawing
relative to the control drawings. In contrast, the Croatian soldiers would be drawn larger and
nearer compared to the control figures. It was expected that these effects would be more
pronounced in children who had lost their father. Hence, the importance hypothesis was tested
for both aversive and friendly topics.

Method
Participants: Sixty children aged between 7 and 10 years (mean age = 8 years 4 moths; SD
= 10 months) were selected from primary schools in Zadar, Croatia. There were 23 boys and
37 girls. Thirty children had lost their father during the recent war.

Materials: White A4 paper and pencils were provided for drawing the men and soldiers. In
addition a pre-dawn outline of a human figure in a smaller piece of paper (10cm by 21cm)
was given to each child so that a self-potrait drawing could be completed.

Design and procedure: The two between-subjects independent variables were trauma
(with or without father) and characterization of topic (neutral man; enemy soldier; friendly
soldier). Children in the two trauma groups were matched on age, gender, school class, exam
achievement (school marks) and drawing ability (as assessed by teacher). Allocation to the
three characterization condition was made on the basis that approximately equal numbers in
the two trauma groups were represented in each condition. The dependent variables were the
differences in size and placement of the second figure drawn compared to the first (neutral)
figure drawn (measured in cm). All instructions to the children were given in Croatian. These
instructions were then translated into English for the purposes of this present paper.

Each child was tested individually in a quiet area within the school during February
1997. The children were presented with the pre-drawn outline of a human figure and
instructed, "Here is a picture which I have started. Please finish this drawing so that it looks
like you. You can draw in a face and anything else you want on the drawing to make it look
like you." On completion of the drawing, it was then attached on top of the left-hand side of a
horizontally presented blank A4 page so that the edges of the two pieces of paper were
aligned. The child was instructed to draw a man on the blank side of the page (the
experimenter pointed to the relevant area). On completion of the "man" drawing the A4 page
was removed and replaced with a blank A4 page that was similarly inserted beneath the self-
portrait drawing.
The instructions for the second drawing differed depending upon what
characterization condition the child had been allocated to. In the control condition the child
was instructed, "Please make a second drawing of the man the same as your first drawing."
Jolly R., Vulić-Prtorić A. – XI. DANI PSIHOLOGIJE U ZADRU, 27.-29. svibnja 1998. 5

For the enemy (friendly) soldier condition the children were told, "recently in our country
there was a war. Please draw a soldier from the enemy (Croatian) army". The blank area of
the page was pointed to in each case.

Results
Two independent raters measured the height and placement of the figures. The height
of the men and soldier drawings were obtained by measuring the distance from the top of the
figure (excluding headgear) to the lowest point of the figure. Placement of the two figures
was mesured by first positioning the self-portrait drawing as it had been originally presented
on each of the pages depicting these two figures.

The distance between the centre of the heads on the self-potrait figure and on the
man/soldier was measured in each case. The raters agreed the size of the figures within 2mm
for 115 out of the 120 drawings. High agreement was also found for placement (111/120).
The measurements used for data analysis were taken from one of the raters except in the
aforementioned instances of disagreement in which a third rater's measurement were used.

The height of the second drawing was subtracted from the height of the first drawing.
The mean difference in height for the two trauma groups by characterization of human figure
is presented in Figure1. Two-way between-subjects ANOVA was calculated to examine the
effects of trauma and characterization of topic on the height of the drawn topic. There was no
main effect for characterization (F2,54 = 0.99, p = 0.38, MSE = 2.39) nor trauma (F1, 54 =
0.08, p = 0.78). There was a tendency for the "without father" group to draw both
characterised soldiers (friendly/enemy) relatively smaller compared to their neutral man
drawings. In contrast, the "with father" group tended to draw both characterised drawings
larger compared to their neutral man drawings. However, the trauma X characterization
interaction failed to reach significance (F2,54 = 2.47, p = 0.09).

The distance of the second drawing from the self-potrait figure was subtracted from
the distance of the first drawing from the self-potrait figure. The mean distance in placement
for the two trauma groups by characterization of human figure is presented in Figure 2. A
two-way between-subjects ANOVA was calculated to examine the effects of trauma and
characterization of topic on the placement of the drawn topic. There was no main effect for
characterization (F2,54 = 0.25, p = 0.98, MSE = 1.48) nor trauma (F1, 54 = 0.36, p = 0.55).
The trauma X characterization interaction was also not significant (F2,54 = 1.95, p = 0.15).

Discussion
The present study's findings indicated that children's feelings towards an emotive topic
did not influence their drawings of the topic's size and placement on the page. Indeed, the
mean differences for both size and placement between the first (neutral man) drawing and the
second drawing (Croatian soldier, enemy soldier or neutral man) were typically below 1 cm
for both trauma groups (with or without father). Notwithstanding these negative findings, the
"without father" group did tend to reduce the size of both friendly and enemy soldiers
Jolly R., Vulić-Prtorić A. – XI. DANI PSIHOLOGIJE U ZADRU, 27.-29. svibnja 1998. 6

compared to their first drawing of a neutral man. This decrease is in contrast to the increase in
size of the second neutral man drawn in the control condition by other children who had also
lost their father. It is possible that the bereavement of losing one's father as a consequence of
the war had a broad negative emotional effect on drawings of soldiers, regardless of what side
they were fighting for. It is usual during bereavement for negative feelings to transgress
towards a number of areas of the bereaved's life. Hence, it is possible that if size is an
emotional indicator then related topics might also be affected.

Recent experimental studies testing the importance hypothesis have used an analogue
procedure in which the emotional significance of a topic is potrayed to "normal" children
through character descriptions of the topic. Although these studies reported positive findings
(Fox & Thomas, 1990; Thomas, Chaigne & Fox, 1989), the lack of further published work in
the area casts doubt of whether reliable effects can be found using this procedure (for
discussion on this issue, see Jolley, 1995; Thomas & Jolley, in press). We believe that for the
importance hypothesis to be properly tested it is essential that there is a personal, emotional
link between the drawer and the topic drawn.

The present study goes someway, therefore, to testing the validity of size and
placement effects by asking children to draw a topic that has represented an important
emotional figure in their lives. Nevertheless, the chosen topic was generic to the children's
emotional experience (i.e., not of any specific soldier). Reliable effects may be more likely to
be shown in children's drawings for particular emotive topics known to the child. For
example, drawings of a sex abuser compared to a non-emotive figure of the same size in real
life might produce consistent size and placement effects in a sample of sexually abused
children where the abuser in known to each child.

We suspect that there is no consistent relationship between emotional significance and


particular features in drawings even in cases where the topic drawn is someone personally
known to the child. That is not to say that children do not express their attitudes towards
topics within their drawings. In fact, for some children we do consider that drawings do
illustrate such feelings. However, we believe that the portraying of emotional importance in
topics is idiosyncratic and unpredictable due to a number of contrasting influences that
emotional significance has on the drawn topic.

Thomas & Jolley (in press) suggest a number of emotional mechanisms of influence
towards the size in which a topic is drawn. One possibility is the aforementioned defence
mechanism in which a child draws an aversive topic relatively small to reduce the perceived
threat of that topic. In addition, Thomas & Jolley suggest that a
child may draw a topic smaller to administer punishment to that topic. For instance, a child
may draw an irritating brother or sister relatively small on one occasion in order to show the
lack of esteem in which the sibling is currently held by the child. Furthermore, a person that
represents an unimportant figure in the child's life may often be drawn small. A similar
mechanism would predict important emotional figures in the family to be drawn relatively
large by our young artist. Payne (1990) reported that in a sample of single mother households
children drew larger female than male figures (see also Issacs & Levine, 1984). This is not
only in contrast to the likely difference in absolute sizes of the parents but also how they are
drawn by children living with both parents (e.g., O'Brien & Patton, 1974).
Jolly R., Vulić-Prtorić A. – XI. DANI PSIHOLOGIJE U ZADRU, 27.-29. svibnja 1998. 7

This mechanism can be extended to socially important topics. Aronsson & Andersson
(1996) have noted that children who experience traditional teaching practices that emphasize
obedience and respect draw reachers larger than drawings made by children in classes that
underplay such respect. Finally, figures may also be drawn relatively large to portray the
power of the topic. In this way threatening topics may be drawn large to depict the power of
the threat that they can exert. Indeed, Jolley (1995) found that children select the larger of two
otherwise identical figures as that representing a nasty character.

Any one of these mechanisms of influence may activate within any one drawing
session, but not necessarily extend to other occasions in which the topic is drawn. Interpreting
which mechanism is active at any one time becomes more confusing when one considers the
production difficulties in drawing. Freeman (1980) has long since alerted researchers to the
influences of planning, available space and level of detail on the size of a drawn topic. Indeed,
these claims have been substantiated by empirical data (e.g., Tsalimi & Thomas, 1988;
Henderson & Thomas, 1990).

So where does this leave drawings as a tool for interpretating children's emotions? It is
clear that they should not be used for diagnostic purposes to assess either the child's
personality or current emotional state. Both positions rely on the dubious body-image
assumption that children project their feelings about themselves in their drawings. The
importance hypothesis seems to us to have wreater face validity as conveying one's feelings
towards a topic in a drawing is less likely to rely on unconscious processes. However, the
relationship between certain graphic features such as size and placement of emotionally
significant topics has yet to be shown convincingly.

We propose that future research tests the validity of interpretating the child's feelings
into their drawings by following two routes. The first attempts to answer the more general
question of whether drawings inform the clinician about the child's emotional disturbance to a
degree beyond which can be obtained by other measures. The second tests the validity of the
importance hypothesis. Both lines of research lend themselves easily to empirical studies.

To answer the first question, we take up the vignette suggestion made by Thomas &
Jolley (in press). Judges with clinical experience would be given case history descriptions or
other forms of assessment of a sample of child referrals. Half of the judges would be provided
also with a selection of drawings made by each child. All judges are then required to
categorise each child into the relevant clinical categories (abused, phobic, etc.). Such an
approach would test whether the drawings significantly increased the accuracy of the judges'
categorization.

In response to the second question, we advise that each of the many mechanisms that
potentially affect size and placement of topic is tested separately. For instance, the defence
mechanism could be tested by selecting a sample of children who have been clinically
assessed for their fear of a particular topic. The drawings of this topic and of a similar life-
sized but non-feared topic could be examined for size and placement effects relative to a self-
portrait figure. Testing the punishment mechanism may pose problems in obtaining an
appropriate sample, but is not beyond feasibility. For instance, children known to be currently
punishing a person (e.g., bullying) could be asked to draw that person, as well as another
Jolly R., Vulić-Prtorić A. – XI. DANI PSIHOLOGIJE U ZADRU, 27.-29. svibnja 1998. 8

important person to the child to which he or she has no similar aversive feelings. Only by
unpacking each mechanism can we gain an understanding of how children portray
emotionally important topics.

Even if both approaches prove to be unfruitful, we would still recommend considering


asking a child to draw during either the clinical interview or subsequent therapy sessions.
Most children up to around the age of 10 year find drawing an enjoyable activity and require
little encouragement to draw. Hence, a drawing activity may aid in calming the child in what
is otherwise a stressful situation, particularly for younger children. If the child gains
confidence during this activity then this in turn is likely to facilitate some bonding between
the child and clinician/therapist. Once the child begins to feel at ease with the adult then
communication between the two may follow. Although the literature to date does not lead us
to recommend applying an interpretation to the child's drawing, it is conceivable that children
may make comments of interest during the drawing process.

For instance, Butler, Gross & Hayne (1995) have shown that 5- to 6-year-oldsO recall
for an event (visit to a fire station) was facilitated by asking them to draw that event compared
to using only an interview technique. Drawing did not facilitate a younger age group (3- to 4-
year-olds). This line of research is certainly worth pursuing. For instance, improvement in
recall could be related to not only the child's langauge skills but also to his or her
representational skills. The lack of either skill may hinder the use of the "draw-and-tell"
approach for young children. Nevertheless further research using drawings in this way may
confirm that drawings are a useful communicative tool for emotional events (Pynoos & Eth,
1986). It would not surprise us if it is the accompanying language and not the drawings
themselves that provide the clinician with important insights into the child's disturbance.

You might also like