Directed forgetting effects in explicit and implicit memory.

Abstract

A retrieval disruption construal of suppression during directed forgetting (DF) was examined. It was argued that it is often difficult to distinguish between retrieval disruption, response disruption, and differential rehearsal interpretations of DF effects. To circumvent the problem of response disruption, two types of tests--exclusive (i.e., interference) and implicit--were used. To circumvent the problem of differential rehearsal, incidentally studied items and a list segregation instruction were utilized. Intraserial cuing (i.e., cuing by item sets) was used in all four experiments. In Experiments 1 and 2, the exclusive test was stem-cued recall and the implicit test was stem completion. In Experiment 1, unique stems (i.e., stems that did not share words across study lists) were used. A DF effect was found only in the exclusive test and only with intentionally studied items. It was argued that a methodological artifact obscured the DF effect in the incidentally studied items. In Experiment 2, repeated stems (i.e., stems that shared words across study lists) were used. The use of repeated stems allowed intrusion rates to be measured. A DF effect was again found only in the exclusive test, but this time in both intentionally and incidentally studied items. In Experiment 3, only a free recall exclusive test was used. A DF effect was found only in the intrusions. Again, it was possible that a methodological artifact obscured the DF effect in the exclusive test. In Experiment 4, Jacoby's (1991) exclusive recognition test was utilized. One half of the subjects was instructed to emphasize accuracy and the other half was not so instructed. A speed-accuracy tradeoff was observed such that the uninstructed group was faster, but displayed a relatively large number of "new"-old errors. In the uninstructed group, a DF effect was observed, but this effect was reversed across semantic and structural study conditions. A DF effect was not found in the instructed group. These findings were interpreted as support for the hypothesis that retrieval disruption affects the use of contextual information normally associated with target information in memory (e.g., Kihlstrom & Evans, 1979).