| Abstract: | Previous research has shown that witnesses who are instructed to recall a crime scenario in reverse order as well as in forward order retrieve more total information than witnesses who recall in forward order twice. The present experiment evaluated the hypothesis that forward recall consists of schema-based retrieval whereas reverse-order recall consists of more datadriven, non-schematic retrieval. It was predicted that more actions consistent with a scenario's schema would be recalled in forward order than in reverse order, more actions incidental to the schema would be recalled in reverse order than forward order, and more schema-based intrusions would be generated in forward order. The results confirmed these predictions. Implications for witness interviewing are discussed. |
