The Wisconsin Card Sorting Test (WCST) is a multifactorial and complex test, and it involves so many different kinds of functions that it is difficult to understand why patients fail. Capacity of 'working memory' is possibly involved in the WCST performance and is considered a relevant factor responsible for the schizophrenics' poor performance. The present study was specifically designed to assess the relationship between 'working memory' measurements and WCST performances of schizophrenics. Furthermore, we investigated the relationship between the cognitive dysfunction and the clinical symptomatology. The following tests were administered to 30 schizophrenics and 25 healthy subjects: WCST, Digit Span Test (Backward and Forward), Digit Symbol Substitution Test and Visuo-Spatial 'working memory' Test, a card test appropriately devised. Clinical assessment included the Italian version of the scale of Krawiecka Manchester Scale (K-MS) and the Outcome scale by Strauss and Carpenter (1972). The 30 patients differed significantly in all the neuropsychological variables from the controls. WCST indexes did not correlate significantly with any of the 'working memory' measures (visuo-spatial and verbal) in the samples studied. No relationship was seen between the neuropsychological performances and clinical symptomatology as evaluated by the K-MS scale. WCST indexes and DSST significantly correlated with the outcome measure. The results do not support the hypothesis that executive dysfunction as evaluated by WCST is attributed to 'working memory' impairment, rather they could suggest that these two neuropsychological functions identify different neurocognitive constructs.