Developmental changes in free recall and serial learning of categorically structured lists

Abstract

Free recall performance with both constant and varied input orders was compared to serial recall performance across three grade levels. Categorized lists composed of frequent instances from categories familiar to young children were presented under both blocked and nonblocked conditions. Results revealed high levels of recall performance and semantic clustering with 'no evidence of production deficiency. Both the amount recalled and the number of instances per category increased with increasing grade level, while the degree of clustering did not. :Seruu ana tree recall performance did not differ due to the high use of semantic clustering in free recall, which made the previously facilitative provision of :m ordering strategy during serial learning no longer useful. Developmental assessments of free recall (FR) learning have generally reported relatively low overall performance levels and corresponding low organizational scores for clustering and/or subjective organization (e.g., Laurence, 1966, 1967). Jablonski (1974) has suggested that organization in the FR learning of children should be diminished to the degree that a large verbal repertoire has not yet developed. Additionally, Nelson (1969) has postulated fhat poor performance on categorized lists may result from children's conceptual categories containing instances which differ from those in adult categories. An alternative, but compatible, explanation of the low recall and organizational levels has been offered by Flavell (1970) with the formulation of a production deticiency hypothesis. The latter proposes that the poor performance is due to a failure to produce semantic mediators during storage and/or retrieval. Such a failure may be due to the reduced development of the verbal repertoire and/or the differences in adult and children's ' conceptual hierarchies. However, a production deficiency need not depend upon either factor. The production deficiency explanation of FR also fits well with developmental differences in serial and FR performance. Posnansky (l974a) has noted that children exhibit high performance levels during serial learning (SL) as compared to typical FR. For example, Posnansky showed that children as young as first graders could master a 9·item serial list in an average of six. trials, while in Nelson's (1969) free·recall study, second·grade