The pictorial superiority effect in recognition memory

Abstract

To test the hypothesis, recently advanced by Nelson, Reed, and Walling (1976), that pictures are more memorable than words because of a superior sensory rather than conceptual code, concept memory was tested so that stimulus form could be divorced from stimulus meaning. In a recognition memory task, subjects in two experimental conditions were instructed to identify test items as “old” only if they appeared in the opposite form from their study form (i.e., study pictures tested as words and study words tested as pictures), and to identify as “new” previously shown items tested in the same form and completely new items. Memory of old concepts was identical for experimental groups compared to control groups, who were tested with same-form items, and there was no difference between groups instructed prior to study about the changed-form tests and those instructed only prior to test. Thus, changing the rule for identifying old and new items had no effect on recognition memory, nor did providing an opportunity for differential encoding by prior instructions. The present results suggest that concepts represented as pictures are remembered better than concepts represented as words, regardless of how such memory is tested, and thus that the pictorial superiority effect stems from superior meaning codes as well as superior sensory codes.