Construct Validity of the Three Motor-Reduced Subscales of the Developmental Test of Visual Perception — Adolescent and Adult (DTVP-A): A Rasch Analysis Model Evaluation

Abstract

Background: The Developmental Test of Visual Perception — Adolescent and Adult (DTVP-A) is a test used by therapists even though its construct validity has not been thoroughly assessed. Purpose: This study evaluated the scalability/interval level measurement, unidimensionality, lack of differential item functioning (DIF) and hierarchical ordering of the items of the three motor-free subscales (Figure-Ground, Form Constancy and Visual Closure) of the DTVP-A. Method: The motor-free visual perceptual performance scores from a sample of 221 participants aged 20+ years (49 adults with and 172 adults without neurological impairment) were used to complete a Rasch analysis of the DTVP-A's three motor-free subscales. Results: On the DTVP-A Figure-Ground (FG) subscale, items FG2 and FG6 demonstrated Rasch model misfit, while no item exhibited DIF. On the Form Constancy (FC) subscale, items FC1, FC2, FC4 and FC6 were misfitting, while items FC3 and FC11 exhibited DIF based on gender. Visual Closure (VC) subscale items VC2, VC3, VC13 and VC15 exhibited misfit with the Rasch model requirements, while VC1 and VC5 demonstrated DIF. When all three subscales were analysed together as one Total Motor-Reduced Visual Perception Scale, items FC2, FC1, FC4, VC12, VC15, FC17 and FG6 demonstrated Rasch model misfit, while items VC1, VC2, VC5 and FC11 exhibited DIF based on gender. Conclusion: The construct validity, scalability, hierarchical ordering and lack of DIF requirements were met by the majority of the motor-free DTVP-A items. However, items FG6, FC1, FC2, FC4 and FC11 appeared to be particularly problematic. Given the fact that a number of items exhibited Rasch analysis model misfit or DIF, clinicians need to take this into consideration when using the three motor-free DTVP-A subscales individually or collectively to assess clients.