Transfer performance in children's oddity learning as a function of dimensional preference, shift paradigm and overtraining

Abstract

Eighty kindergarten and second grade children were trained on an oddity problem with their preferred dimension either relevant or variable and irrelevant. After reaching criterion, half the Ss received 50 overtraining trials prior to the administration of transfer problems while the remaining Ss were transferred immediately to the new problems. Transfer conditions introduced all new stimuli and Ss were randomly assigned to either an intradimensional (ID) oddity shift or an extradimensional (ED) oddity shift. Dimensional preference had a significant effect on original learning, but no effects of this variable were revealed in the transfer data. The older Ss acquired the original problem in fewer trials. For both age groups, ID transfer shifts were learned significantly faster than ED shifts following a weak criterion measure. For the kindergarten Ss overtraining slightly but not significantly increased the difference between the shift conditions. However, both the ID and ED shift performance of the older Ss was facilitated by overtraining and near perfect transfer scores were recorded. An explanation was put forward in terms of chaining theories of learning which stress mediation through covert attentional responses.