The validity of children's human figure drawings for identifying psychopathology in a community sample of children

Mary Bruce Webb, Fordham University


Although surveys show that children's Human Figure Drawings (HFDs) are among the most commonly used personality assessment measures, few studies exist investigating the validity of the HFD for identifying young children with emotional disorders. This study examined the relationship between the HFD and other assessments of psychopathology in a community sample of children ages 4 to 8 years, specifically (a) the Physician's Rating Scale (PRS), (b) the Child Behavior Checklist (CBCL), and (c) the Teacher Report Form (TRF).^ The CBCL and PRS were administered in pediatricians' offices for all children in this age group (N = 1886) who visited the offices during selected periods. CBCL and PRS scores were used to classify children into clinical and comparison groups. All children in the clinical group, and a sample of children in the comparison group, were selected for follow-up assessments, which included the HFD and TRF. HFDs were obtained for 1059 children and scored for Emotional Indicators (EIs) according to Koppitz's instructions.^ Few individual EIs significantly discriminated clinical from normal groups. Clinical group children were significantly more likely to have 2 or more EIs in their drawings, but over one half the comparison group also met this criteria. A stronger discriminator was a combination of having 2 or more EIs, and the inclusion of certain key EIs. Results raised doubts about the utility of HFDs as a screening instrument for psychopathology, since large numbers of children could be misclassified on the basis of HFD results.^ Kappa coefficients were computed to determine if the HFD classified the same children as clinical "cases" as did the CBCL, TRF, and PRS. Kappa coefficients also were computed for subgroups divided by age and sex, and by CBCL and TRF subscale classifications. Kappa coefficients were uniformly poor for the HFD compared with the other instruments. However, kappa coefficients also were poor when the criterion instruments were compared with each other. The study highlighted the need to use multiple technologies and informants in assessing emotional functioning. ^

Subject Area

Educational tests & measurements|Psychology

Recommended Citation

Webb, Mary Bruce, "The validity of children's human figure drawings for identifying psychopathology in a community sample of children" (1993). ETD Collection for Fordham University. AAI9412162.