Effects of Conceptual Grouping of Adjacent Items on Serial Learning

Abstract

10 Ss were randomly assigned to each of the 6 experimental conditions representing placement of one of two 5-item concept clusters, empirically shown to differ in degree of conceptual relatedness, in either the beginning, middle, or end portion of a 15-item serial list. Ss learning lists containing a highly conceptually related cluster exhibited a greater tendency to learn the concept items early in serial learning, regardless of their position in the list, than did Ss learning lists containing a less conceptually related cluster. As a result of a superiority of performance in the cluster portions of the high-concept lists and an opposite superiority of performance in the non-cluster portions of the low-concept lists, the serial position curves for the 6 lists were significantly different. The failure to find over-all differences in learning among lists yielding different serial position curves supports theories of serial learning which propose that total serial learning time is independent of the order in which the items of a serial list are learned.