Prediction of Psychiatric Hospitalization, Diagnoses, Arrests, and Violent Behavior Through Scored Drawings and Associations [2017]

The article presents a completely empirical approach to using drawings and associations as a supplemental tool in forensic psychological assessment. The present system represents a modern alternative to prior theory-driven “projective technique” approaches that have failed to pass scientific scrutiny. Using an archival sample of 252 forensic cases, House-Tree-Person drawing protocols were scored according to preliminary test items developed from published handbooks, academic courses, and workshops on the use of drawings and related associations for assessment. Intermediate test items were selected according to their being significantly correlated with at least one external psychometric, behavioral, or clinical variable. Psychometric and editorial considerations guided the development of a final set of 66 test items. A “flagged for pathology” approach demonstrated modest convergent and discriminant validity for the final set of test items against clinical criteria on an independent sample of 150 cases of the same type used in the preliminary item development phase. Included are sample items and discussion of psychometric issues in working with this class of data.

Frank J. Dyer

Psychological Reports, Vol 121, Issue 1, 2018