Construct validity of the Wisconsin Card Sorting Test-64 in patients with stroke.

This study examined the factor structure and contrasted-group validity of the Wisconsin Card Sorting Test-64 (WCST-64) in a stroke sample (n = 112). Confirmatory factor analyses were used to compare five different models suggested by prior factor analyses. The results indicated that the WCST-64 was best represented by a three-dimensional model comprising response inflexibility (factor 1), ineffective hypothesis-testing strategy (factor 2), and set maintenance (factor 3). A significant overall multivariate effect for group (F = 2.87, df = 18,495.46, p <.001) was found in a multivariate analysis of covariance with WCST scores as dependent variables and four different groups (three stroke subgroups with different levels of cognitive function and a normal control group) as independent variable, after controlling for gender. The results of discriminant analysis supported the use of the WCST-64 in stroke patients with cognitive impairment.