Abstract
<p>False memory (FM) has been widely studied using the Deese-Roediger-McDermott (DRM) paradigm. Across two experiments, the present work examined the roles of emotional valence, gender, and working memory capacity (WMC) in false and correct recall and recognition. Experiment 1 tested the prediction that women would show greater FM for emotionally valenced lists than men. Contrary to predictions, the gender x valence interactions were nonsignificant for both false recall and false recognition. Experiment 2 was a planned replication and extension, incorporating WMC measures (RSPAN, SymSpan, and backward digit span) to test whether higher WMC would be associated with reduced FM and to explore whether this association differed by gender. Higher WMC was associated with greater correct recall and recognition, and with lower false recall and recognition. Women showed greater correct recall and recognition in Experiment 2, but the predicted gender differences in FM did not emerge. Exploratory treatment-coded Gender x WMC and Gender x Valence x WMC coefficients were statistically significant for false recall, although the corresponding omnibus interactions were nonsignificant and the patterns were not in the predicted direction. Across both experiments, negative lists generally produced the greatest false recognition, whereas false recall was generally lower for positive than negative lists. Overall, emotional valence affected recall and recognition differently, and higher WMC was consistently associated with greater memory accuracy.</p>