Construct Validity of the Wisconsin Card Sorting Test-64 in Patients with Stroke

Abstract

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.

Keywords:

• Factor structure
• Stroke
• Validity
• Wisconsin Card Sorting Test-64

ACKNOWLEDGMENTS

This work was supported by a grant from the National Science Council (NSC) (project NSC 87-2314-B-037-012). We are most grateful to study participants for their contribution to this study.

Notes

NI = group with no cognitive impairment; MI = group with mild cognitive impairment; MSI = group with moderate to severe cognitive impairment; GCS = Glasgow Coma Scale; RBD = right brain damage; MMSE = Mini-Mental State Examination; Values are expressed as means with standard deviations in parentheses.

M₁ = one-factor model; M₂ = two-factor model; M₃ = Greve et al. (2005) three-factor model; M₄ = M₃ with FMS, CLR, CAT, and TC loading on the set maintenance factor; M₅ = M₄ with three factors covarying with each other; NNFI = Non-Normed Fit Index; CFI = Comparative Fit Index; RMSEA = Root Mean Square Error of Approximation.

TC = total correct; PE = perseverative errors; NPE = nonperseverative errors; CLR = conceptual-level responses; CAT = categories completed; FMS = failure to maintain set.

WCST = Wisconsin Card Sorting Test; ANCOVA = analysis of covariance; SE = standard error; NI = group with no cognitive impairment; MI = group with mild cognitive impairment; MSI = group with moderate to severe cognitive impairment; TC = total correct; PE = perseverative errors; NPE = nonperseverative errors; CLR = conceptual level responses; CAT = categories completed; FMS = failure to maintain set.

NI = group with no cognitive impairment; MI = group with mild cognitive impairment; MSI = group with moderate to severe cognitive impairment; PR = perseverative responses; CLR = conceptual level responses; TC = total correct; NPE = nonperseverative errors; CAT = categories completed; FMS = failure to maintain set.

^a p < .01.