Validity and norms for the Wisconsin Card Sorting Test (WCST) in its computerized and manual versions in children

The Wisconsin Card Sorting Test (WCST) was initially developed to measure problem solving and abstract reasoning. Currently, it is widely used for the neuropsychological evaluation of those cognitive processes related to frontal lobe functioning, more specifically as a measure of reactive cognitive flexibility. Although its use is widespread both in clinic and research settings, controversies remained regarding the nature of the cognitive processes that underlie WCST performance and the relationships between its diverse scores and other related cognitive measures. To date, studies have been conducted in adolescents and adults, where two- and three-factor solutions predominate, although a one-dimensional structure has also been reported. On the other hand, there are few normative studies in the Spanish-speaking population, and, to our knowledge, there are no normative studies that contemplate the two versions of the WCST. The aim of the present study was to examine the factorial and convergent validity of the WCST in its computerized and manual version in children and to establish normative values for each version indicator according to age and sex. The sample consisted of 359 children aged 7 to 12, divided as follows: (1) manual WCST: 138 children aged 7 to 12 from both sexes and (2) computerized WCST: 221 children 7 to 12 from both sexes and. The inclusion criteria were the following: (1) children with no clinical, neurological, or psychiatric history; (2) who attend their school studies regularly; (3) without school repetition. The WCST in its two versions (computerized vs manual), the Kaufman Brief Intelligence Test (K-BIT) to assess children intellectual functioning, and the working memory (WM) subtask of the WISC-IV, the Stroop Color and Word test and the semantic and phonological verbal fluency (VF) test were administered to test the WCST convergent validity. Intellectual functioning (IQ) was within the normal range expected for the age of children who completed both versions of the WCST. Exploratory Factor Analysis (EFA), Pearson's bivariate correlations and multivariate analysis of variance (MANOVA) were used. For both the computerized and the manual WCST, the parallel analysis yielded a one-dimensional solution, grouping efficient and inefficient responses, with Complete Categories (CC) being the indicator that obtained the most weight. When comparing the structure found for each version, it is observed that Total Correct (TC) weighed only in the computerized version of the WCST. Regarding the WCST convergent validity, the results with the computerized WCST indicated positive and significant relationships (r r = .17 to r = .33) between Total Correct and Complete Categories of the WCST and IQ, MT, FV and Inhibition. In turn, Perseverative Responses, Perseverative Errors, and Non-Perseverative Errors, presented negative and low correlations (r r = .17 to r = .29) with IQ, MT, and FV. For the manual WCST, only Complete Categories presented significant positive correlations (r r = .20 to r = .39) with CI, FV, Inhibition and MT. The other indicators of the manual WCST presented negative correlations (r r = .18 to r = .44) with CI, FV and MT. Regarding the factorial validity of the instrument, our results suggest that the structure of the WCST would be unidimensional in children, regardless of the type of version under analysis (i.e., manual vs. computerized). However, the relationship between WCST performance and the tasks that assess other cognitive processes differs depending on the version. Normative data is established for the first time for Argentine children for both versions. The results emphasize the importance of not considering both versions as interchangeable in child populations and interpreting each version according to its respectively norms.