Some factors affecting discrimination learning in young children

Abstract

An experiment was designed to test the effects of certain factors on the discrimination learning and shift performance of young children. Subjects were selected between the ages since it has been observed that at about that age, reversal and non-reversal shifts in discrimination are executed equally readily. Five dichotomous variables were controlled: intelligence, sex, social group, type of shift, and reward condition. Subjects were assigned on their performance of the childrens' Progressive Matrices, to high or normal IQ groups. Males and females were separately grouped. Social group was determined by school attendance: tv/o schools in middle class residential districts, and two in working class areas were visited. Children were allocated at random to shift and reward conditions. After initial training, half the children were rewarded for performing a reversal or intradimensional shift, the other half, a non-reversal or extradimentional shift. Half of the subjects received material incentives: sweets or trinkets for correct responses, and half, nonmaterial rewards: bell tinkles or light flashes. The performance of 128 children was examined by five way analyses of variance. The variables' effects were minimal during initial training, but in the discrimination3shifts, interesting effects emerged. The group as a whole performed both shifts with relative ease, but middle class children performed reversal shifts better than non-reversal ones while working class children performed non-reversal shifts best. Reward effects interacted with intelligence: high IQ subjects performed best for material incentives. Girls generally performed better than boys, although the effects of sex interacted with both IQ and social group membership. A strong relation appeared between verbal facility as measured by the WISC Vocabulary and shift performance. A covariance analysis, by equalizing the impact of fluency, reduced the variability within groups, and accentuated the previously observed effects. A reward choice technique produced evidence of a developing preference for larger, delayed over smaller, immediate rewards in middle class children. The results of this study indicate that the factors selected for study here do affect the discrimination learning of six year olds.