Trade-Off Between Memory for Verbal Items and Their Visual Attributes

In an experiment with 96 high school students, instructions to attend to the case and color in which words were presented led to improved retention of these visual attributes but depressed recognition performance. This trade-off between item and attribute memory occurred at 3 presentation rates (1.5, 5.5, and 10.5 sec/item), suggesting that visual attributes of verbal stimuli are rarely stored unless Ss expect that knowledge of physical form will be useful later. Visual attributes of high-imagery words were retained better than visual attributes of low-imagery words by Ss instructed to attend to case and color, but not by Ss asked only to attend to word meaning. Implications for the hypothesis that visually presented verbal stimuli are stored as literal copies are discussed