Human Figure Drawing Test.

Review of the Human Figure Drawing Test
Charles A. Peterson
Buros Center for Testing

Projective drawings have long been part of the psychodiagnostician’s repertoire. Surveys too numerous to mention document the procedure’s robust popularity, one that has been relatively stable (never left the top 10) throughout clinical psychology’s lifetime.

For better or for worse (yes, it seem like we are wedded to these procedures), human figure drawings rest on several assumptions: (a) Frank’s (1939) classic statement of the projective hypothesis argues that the psychological structure of the individual will express itself as it organizes unstructured material (cf., Peterson & Johanson, 1993; Peterson & Schilling, 1983 on the classification issue). (b) The choice of the human person/body for graphomotor expression is a deft one because the ego is “first and foremost a body-ego” (Freud, 1923, p. 27). (c) The procedure achieves its projective/unstructured status with a simple instruction, one that asks the test subject to “draw a person,” assiduously avoiding cues or rules that might guide or structure the production. (d) Finally, when the test subject draws a human form, the subject’s personality structure and dynamics will be revealed to the psychologist for diagnostic scrutiny.

It may be that the unstructured nature of the task is useful for the test subject, as it facilitates the expression of personal material for the divination process. But, it is equally fair to note that the unstructured nature of the task is not good for the psychometrically minded psychologist.

The Human Figure Drawing Test (HFDT) is one more attempt to organize and systematize the interpretation of the unstructured projective procedure. The authors insist that their procedures will add some quantitative data to enhance–not replace–the more typical free-form qualitative analysis. This version of the Draw-A-Person provides a “Drawing Form” upon which demographic information is collated, and upon which the test subject draws a person. The clinician is encouraged to note unique reactions and differential interests as the subject does so. Using criteria illustrated in the manual, the same-sex drawing is scored according to 74 criteria, such as “elaborate belt or waist emphasis” or “chicken feet” (talons or multiple lines) on a convenient form, which has been cleverly backed with carbon paper, transferring scores to the “Interpretation Guide.” In an amazing feat of prestidigitation, the scores yield written interpretations next to the score checked. For example, “waist emphasis” equals “concern over sexual control,” and “chicken feet” equals “regression or retardation,” and “disheveled hair” equals “sexually impulsive behavior” or “thought disorder.” Certain scores are noted for “simplification” and other scores are tallied on a “distortion” index and later arithmetically manipulated to produce an “Organic Factors Index.” References are provided for the interpretation of such factors as “thin, wasted or ribbonlike arms.”

The test data are interpreted on multiple levels, several of which produce indices related to cognitive impairment: (a) The Impairment Scale provides information on whether or not the individual shows some cognitive impairment. The authors caution “the clinician must remember that this test is designed to detect only general variations in the level of cognitive functioning” (manual, p. 71). Translate: this test is not that accurate. Use a Wechsler and consult a neuropsychologist. Later on, Table 3 presents “features believed to represent organic impairment” (manual, p. 159; emphasis added). In other words, they present no data to indicate that such factors are related to organic impairment. It is also puzzling to be told that the same 74 indicators that were used to divine functional psychopathology are now the same factors believed to represent organic impairment. Further, the algorithm underlying these 74 factors uncritically assigns a weight of 1 to each indictor, many, no doubt, of differential predictive validity. (b) The test procedure also yields “Distortion” and “Simplification” scores, used in a very fuzzy manner to rarify the nature or severity of the subject’s cognitive deficit as measured by the “Impairment Scale.” (c) An “Organic Factors” score is obtained by subtracting the Simplification from the Distortion score. A range of scores indicates, at the antipodes of the scale, either an “organic condition” or “mental retardation” whereas scores in the middle neither confirm nor exclude the presence of either condition. (d) Finally, anchored by the above “quantitative” interpretations, the test results are now sifted for qualitative interpretation of such factors as “bare feet on a clothed figure” (equals “aggression” for the curious reader).

After 16 case examples, the reader arrives at a chapter on “development and psychometric properties” (manual, p. 159). Seven hundred subjects from a state hospital composed the clinical sample, and the nonclinical group was made of 100 hospital staffers not screened for psychopathology. Differences in age, race, education, and intelligence were noted among clinical and nonclinical samples, but the differences were not subjected to tests of significance. Each drawing was scored by three psychologists, and a two out of three agreement was required to score the presence of a particular sign such as “nostrils showing.” Interrater reliability scores are not provided for individual items. Further, the manual is quite confusing on who is agreeing with whom, with both inexperienced raters and more experienced raters (psychologists) producing Cohen’s kappas of .77, .42, .57, and .23, none-too-impressive numbers for the major Impairment index. The various features of the drawings are tabulated in frequencies for each diagnostic group. It is not clear how these diagnoses were obtained. The manual offers some data to suggest that Impairment Scores are related more to diagnosis than IQ (not stating how the latter was obtained). Items on the Distortion and Simplification scales were factor analyzed and revised to discard low-weighted items and to enhance Cronbach’s alpha. The two scales correlate .56 with one another, indicating considerable overlap. No test-retest data are provided. Each of the scales (i.e., Distortion, Simplification, total Impairment scale, and the Organic Factors Index) were able in discriminant analyses to differentiate among nonclinical, psychiatric, and organic/mentally retarded groups at a rate exceeding chance classification. Sensitivity and specificity data should have been provided here.

SUMMARY. Human figure drawings–for all their problems–continue to be used by a majority of clinical psychologists. The authors of this version of the draw-a-person may be commended for their effort to structure the interpretation of this unstructured technique. In general, there is more support for the composite indices than for individual items, here repeating Swenson’s (1957) findings and cautions to focus on broad, robust factors such as adjustment–maladjustment. Whether it is worth going through all this to conclude a patient is maladjusted is open to serious doubt. In the end we may recall, but question, Freud’s words (a February 28, 1934 letter to projective theorist Saul Rosenzweig): This type of research will convince no one, but “can do no harm” (quoted in Shakow & Rapaport, 1964). However, fuzzy research can do harm. If one makes the precarious decision to use figure drawings, then it could be argued that one should use a version that imposes structure on the diagnostician’s divinations. Partisans on both sides of the aisle will no doubt continue to project their sketch of the utility of human figure drawings. Still, and staying within the metaphor, fancy raiment on a stick figure does not really improve the stick figure.

REVIEWER’S REFERENCES

• Freud, S. (1923/1961). The ego and the id. In J. Strachey (Ed. & Trans), The standard edition of the complete psychological works of Sigmund Freud (vol. 19, 3-66). London: Hogarth (originally published 1923).
• Frank, L. K. (1939). Projective methods for the study of personality. Journal of Psychology, 8, 389-413.
• Swenson, C. H., Jr. (1957). Empirical evaluations of human figure drawings. Psychological Bulletin, 54, 431-466.
• Shakow, D., & Rapaport, D. (1964). The influence of Freud on American Psychology. Psychological Issues, Monograph 13 (whole issue).
• Peterson, C. A., & Schilling, K. M. (1983). Card pull in projective testing. Journal of Personality Assessment, 47, 265-275.
• Peterson, C. A., & Johanson, T. J. (1993). A taxonomy of psychological tests. British Journal of Projective Psychology, 38, 63-70.