Fifty years of memory for names and faces: A cross-sectional approach.

SUMMARY Three hundred and ninety two high school graduates were tested for memory of names and portraits of classmates selected from yearbooks. The retention interval since graduation varied from 2 wk to 57 yr. Performance was adjusted by multiple regression procedure to control the effects of various conditions that influence original learning, such as class size, and other conditions that influence rehearsal of the material, such as attendance at class reunions. The cross-sectional approach illustrated in this study is discussed as an alternative to the traditional longitudinal approach to the study of memory. The cross-sectional adjustment approach sacrifices laboratory control over the conditions of original learning, but permits extension of investigation s to conditions that significantly affect learning and memory but that cannot be adequately or conveniently manipulated in the confines of the laboratory. Examples are a greater range of learning material, of motivational conditions, of context effects, and most significantly of acquisition time and retention time. Since most of the vital information in the memory store is learned and retained over periods much longer than those which are investigated in the laboratory, available generalizations based on laboratory findings may be inadequate or inapplicable to these area of knowledge. The cross-sectional approach provides a valuable option for exploring this issue. The data of this study show that identification and matching of names and faces remain approximately 90% correct for at least 15 yr even for members of very large classes. Free recall of names is independent of class size and of recognition memory and declines with negative acceleration by 60% .during 48 yr. This very slow forgetting process is attributed to a great expansion of distribution of practice and overlearning effects. This study also argues that laboratory-defined tasks generally impose a greater interdependence upon free recall and recognition performance than is inherent in the memory system. The fact that recall is uncorrelated with class size or recognition performance supports models of memory in which generative and retrieval processes are independent of storage processes. Social context is shown to be a very important determinant of recall performance, and a much less important determinant of recognition performance. Apparently the context serves as a mediator for retrieval of information; once the association of mediator and target item is lost, the item can no longer be retrieved but can still be identified on a recognition task. •