The Handler Anxiety Index (HAI) was developed to provide a measure of anxiety from the Draw-A-Person (DAP) protocol. The present study examined the dimensionality and concurrent validity of the HAI against 2 criteria, clinical diagnosis and the 16PF Anxiety scales, in 120 16–47 yr old medically or psychiatrically disabled patients. Results are as follows: (a) The median reliability estimate for the 20 HAI items was 0.97; (b) the HAI items were heterogeneous with respect to item–total structure and saturation with drawing quality; (c) factor analysis of the HAI produced 6 orthogonal dimensions: Poor Quality, Correction, Three-Dimensionality, Expansion–Constriction, Overinclusiveness, and Inadequate Control; and (d) there was no pattern of significant relationships in the correlations of the 6 HAI factors with the 2 independent anxiety criteria. Results suggest that there are no subsets of drawing elements that are valid indicators of anxiety. (18 ref)