Preliminary analyses of human-figure drawings as a kindergarten screening measure.

Abstract

Human-figure drawings have been used to identify children with emotional problems successfully (Koppitz, 1968; Landsman & Dillard, 1967). T o date, the best scales include items selected individually by means of empirical findings. Through combining two widely used measures (Koppitz Emotional Indicators and Evanston Early Identification Scale) we hoped to provide a basis for rhe construction of a rational scale. Our spccific aims were ( a ) to determine which items correlated significantly with the combined total scale scores. ( b) to determine which items correlated significantly with independent judgments of clinical high risk, and ( c ) to factor analyze the total item pool in order to determine whether [here were any underlying cohesive factors predictive of clinical high risk. One hundred and ten kindergartners, from diverse socio-economic backgrounds and schools and ranging in age from 4.11 to 6.10 yr., were given the Draw-A-Person test. O f 56 males and 54 females, 62 were Caucasian, 38 were black, 3 were Asian, and 7 were Spanish-surnamed. Following extensive classroom observation plus individual testing when necessary, all were judged to possess normal intelligence by us. Inter-rater agreement on scored drawings war 96.88% for the Evanston and 93.66% for the Koppitz. The Koppitz contains 23 items of equal weight; the Evanston contains 10 weighted items, 6 of which appear in unweighted form on the Koppitz. An independent ranking of clinical risk was provided by graduate students in psychology, following extensive classroom observation. An item analysis of the combined 27 unweighted items yielded five which correlated at .50 or beyond ( p < .01) with the combined totals of both scales: no body. .60; no mouth. .61; no eyes, .58; no nose, .56; poor integration, .50. An interitem correlational matrix yielded the following significant Pearson productmoment correlations between individual items and independent criterion of clinical high risk: ( a ) .36 no body, .27 correct position ( p < , 0 0 1 ) ; ( b ) .25 no mouth ( f i < . 0 1 ) ; and (c ) .23 poor integration, .22 tiny figure, .21 no nose ( p < .05). An exploratory factor analysis was undertaken based on this small sample of 110. All principal components were obtained; 1 1 had roots greater than one and were rotated, using Ka~ser's Varimax Criterion. Only rwo of the rotated factors accounted for more than 28 9; oof the total variance and seemed important conceptually. Clin1c31 judgment of high risk plus omissions of central body parts (mouth, nose, eyes) had high factor loadings on our first factor; peripheral body-part omissions (legs. arms) had high loadings on our second factor. Together, a relationship is suggested between body integrity and severity of problem worth further exploration. REFERENCES KOPPIIZ, E M. Psychological evaluation of children's human figure drawings. New York: Grune & Stranon, 1968. LANDSMAN, M., & DILLARD, H. Euanston Early ldenrification Scale, manual. Chicago: Follett Educational Corp., 1967.