Effects of typicality on recall and clustering in a free recall task: a developmental study

Abstract

This study was designed to explore the effects of semantic typicality of category members upon first and third graders' recall and clustering in a multi-trial free recall task. In Experiment 1, members of 5 superordinate categories were elicited directly from 32 first grade and 33 third grade children using a category item production task. Size, diversity, and agreement among the category items were examined by age and category. Children of both age groups exhibited large, diverse, and non-idiosyncratic category conceptual schemes and no age differences occurred in number, diversity, and extent of agreement among category members first retrieved from long term memory. From among the age-appropriate category members, prototypical and nonprototypical items were selected to form all prototypical, all nonprototypical, and mixed (i.e., prototypical and nonprototypical) recall lists for the three list conditions of the free recall task used in Experiment 2. Sixty children (30 first and 30 third graders of the same ages, background, and school as the children in Experiment 1) were alternately assigned to one of the three age-approriate list conditions of Experiment 2. The items in each list condition were randomly ordered for each of the five free recall trials with the restriction that no two items from the same category were adjacent and that each trial began and ended with different items. Order of presentation of the five randomized lists was constant across children within each list condition. The number of items recalled and the extent of category clustering were examined. There were no significant age differences in recall and clustering with the all prototypical lists and the all nonprototypical lists. First grade children recalled as many items as third grade children and clustered their recall to the same extent as third grade children with these two types of lists. For both ages, the all prototypical list condition resulted in the highest recall and highest clustering of the three list conditions. In the all nonprototypical list condition, neither age group categorically clustered their recall. In the mixed condition, the well established age-related differences in recall and clustering occurred. Third graders recalled significantly more items and tended to have higher cluster scores, and significantly higher scores on the last two trials, than first graders. Examination of repetitions and intrusions in recall within the mixed list condition indicated that the older children used more effective retrieval cues than the younger. children, while the younger children had almost twice as many poor retrieval cues than the older children. By manipulating semantic typicality of age-appropriate' categories and items, age-differences can be either produced or eliminated. These results suggest that: (1) Past free recall studies which reported significant age differences may have used mixed typicality lists and thus may not have tapped the full range of young children's competency; (2) Production deficiency does not appear to be a child characteristic or an age-related characteristic. Rather production deficiency more aptly describes the joint functional outcome of the type of information to be recalled and the ease (or difficulty) of subsequent retrieval demands; and (3) Further work which concentrates on retrieval strategies and their interaction with typicality and age may contribute to our understanding of memory development.