Discussion centered on three sets of factors that may influence the variability observed in "concept identification" (concept formation and discrimination learning) studies: stimulus characteristics, incentives employed, and subject characteristics. Stimulus characteristics are described in terms of the number of dimensions simultaneously present in the conceptual task. A study which investigated developmental differences in concept identification when memory aids were introduced for tasks of varying levels of complexity was described. The aids were helpful to all children, but especially to older subjects, and there was a remarkable variability in performance noticed. The incentive conditions of such studies may account for some of this variability, as indicated by reward versus punishment studies in which the punished subjects performed reliably better on complex tasks than rewarded subjects do. Finally, the effect of subject variables, in terms of cognitive style was investigated through a study of relationships between impulsive style and Stanford Binet I.Q. Thirty-three disadvantaged preschool children were administered two tests of ability to inhibit motor ability. Results showed that the two measures did correlate with the Binet in this sample, and the ability to inhibit motor performance had little to do with understanding of test instructions. Thus, cognitive style may be an important determiner of conceptual identification. (DP)