Human Figure Drawing in a Pakistani Adult Clinical Sample: Base Rates, Demographic Correlates, and Presenting-Problem Differentiation

Background. The Draw-a-Person procedure remains among the most widely administered instruments in clinical practice worldwide, yet almost the entire empirical literature on its indicators derives from Western child samples. No published base rates exist for adult human figure drawings (HFD) in Pakistan, and interpretive cut-points in routine use were established in populations whose drawing conventions, dress, and visual culture differ substantially from those of South Asian clinical attenders. Objectives. To establish base rates for 52 HFD indicators in a Pakistani adult clinical sample, to test whether age, sex, and education predict drawing characteristics, and to determine whether presenting problem differentiates drawings once multiplicity is controlled. Method. One hundred and one consecutive adult clinical attenders (60 female, 40 male, 1 other; M age = 28.53, SD = 8.77) completed a single-figure Draw-a-Person task. Drawings were scored on a 52-indicator protocol spanning structural, completeness, proportionality, special-feature, and global-rating domains. Presenting problems were coded from clinical records into six non-exclusive categories. Analyses comprised base-rate estimation with Wilson confidence intervals, Welch t-tests with Hedges-corrected effect sizes, Fisher exact tests, and multiple regression, with Benjamini–Hochberg control of the false discovery rate across the differentiation analyses. Results. Twenty-one protocols (20.8%) contained no torso, arms, or legs and are described here as a head-only response class; these drawings occupied the page as fully as full figures (d = −0.02) but scored 8.6 points lower on completeness (d = −2.50). Ninety-eight per cent of figures were drawn without a baseline. Hats were absent from every protocol; pockets appeared in 9.9% and buttons in 18.8%, against hair in 91.1%. Age and education were unrelated to figure size, completeness, and global ratings. Women drew substantially smaller figures than men (d = −0.75, p = .0008), an effect that survived adjustment for anxiety, age, and figure type. Anxiety was associated with smaller figures (d = −0.58) and trauma with less complete ones (d = −0.69), but no presenting-problem effect survived false-discovery-rate correction across 25 tests (minimum q = .076). Conclusions. HFD indicators in this population depart sharply from published Western expectation. The floating figure, treated in the classical literature as pathognomonic of insecurity, is here the population norm and carries no discriminative information. A substantial minority of adults respond to the Draw-a-Person instruction by drawing a head rather than a person, which renders proportionality indices undefined and depresses completeness for reasons unrelated to psychopathology. Sex is the only demographic variable that predicts drawing characteristics. Presenting-problem differentiation is at best weak and does not survive control for multiple comparisons. These base rates provide the first normative reference for adult HFD interpretation in Pakistan and argue against the inferential use of individual indicators in clinical and, particularly, forensic contexts.