Abstract
The objective was to examine whether the lower accessibility of studied items (Rp−) that follows retrieval practice with studied items from the same category (Rp+; retrieval-induced forgetting) is correctly monitored by our cognitive system. If monitored, lower confidence for Rp− items would be expected which, in turn, would allow the control of the retrieval-induced forgetting through the report option. In Experiment 1 the standard retrieval-practice paradigm with categorised word lists was followed by a recognition test with confidence rating and the option to report or withhold the answer. Accuracy showed retrieval-induced forgetting, but there were no differences in confidence. The report option did not affect retrieval-induced forgetting. The confidence–accuracy dissociation could be due to a correct monitoring of the retrieval-induced forgetting joined with a factor that incorrectly increases confidence for Rp− items. Familiarity with the practised category was proposed as this factor and tested in Experiment 2. Despite presenting the categories more times during the retrieval-practice phase to increase their familiarity, confidence ratings were unaffected. In conclusion, this research suggests that retrieval-induced forgetting was not monitored, giving rise to a confidence–accuracy dissociation.
We thank Teresa Bajo for helpful comments on an earlier version of this article.
We thank Teresa Bajo for helpful comments on an earlier version of this article.
Notes
1 Retrospective confidence may refer either to a judgement that probes relate to studied items, or to a judgement that the answer is correct. We consider that these two types of judgements are based in the same psychological representation of confidence (cf. Higham, Perfect, & Bruno, Citation2009), and thus use the term in either way.