Investigating young children’s human figure drawings using Rasch analysis

Abstract

The Goodenough–Harris Drawing Test (GHDT) is a non-verbal assessment designed to infer young children’s levels of intellectual development and understanding via the collection of three human figure drawings (HFDs) – one each of a man, a woman and a self-portrait. This paper presents findings from a research project that applied the Rasch model for measurement to HFDs collected from 246 children aged between 4 and 10 years to assess the psychometric properties of the GHDT assessment, and young children’s HFDs in general, as the GHDT had not yet been examined from a modern test theory perspective in full. Results indicated that: (1) the GHDT and children’s HFDs were apt for Rasch analysis and deemed to be generally psychometrically sound; and (2) that children performed almost identically (within error) on all three HFDs collected for the assessment, suggesting that the collection of three HFDs – as well as many of the 217 items which comprise the GHDT – was potentially redundant. Consequently, a prototype human figure drawing continuum (HFDC) was constructed from those data and examined. Rasch analysis revealed that the researcher-constructed 45-item HFDC scale – requiring just one self-selected HFD of a man, woman or child – was just as effective as the 217-item GHDT. The HFDC, when normed, should be seen as a parsimonious and child-friendly HDF assessment.