An experiment was conducted to test the ability of Ss to make logical decisions on the basis of correlated and noncorrelated binary variables. Ss were asked to decide on hospitalization or nonhospitalization for a patient with a disease on the basis of past recovery rates of other patients. It was assumed that only Ss with a cognitive structure similar to that described by Inhelder and Piaget (1958) would behave logically in the experimental task. The findings suggested that such a structure was present in that the Ss were capable of making logical decisions on the basis of the data presented to them. From an analysis of the problem the point of view is presented that previous negative findings in the area may have been a function of methodology, rather than the lack of a cognitive structure as has been claimed.