Download Presentation (567 KB)
This study aimed to compare false memories in short and long-term tests for semantic and phonological lists. The purpose of this study was to replicate our previous results with shorter lists in the short-term memory condition. Both the past and current study used 36 DeeseRoediger-McDermott (DRM, Roediger & McDermott, 1995) lists for the creation of simple false memories. The lists varied in their association to a nonstudied critical lure item: There were 18 semantic and 18 phonological lists. The first study used 6-item DRM lists, whereas the current study used 4-item DRM lists to ensure list length did not exceed working memory capacity. Both studies included recognition tests at short (less than 1 second delay) and long (after all lists had been presented) delays. In the short-term condition, lists were presented one at a time followed by an immediate one-item recognition test. In the LTM condition, after all lists were presented, there was a 1-minute break before a recognition test that tested all lists. The current study replicated the results of the first study, with almost identical mean proportions of false alarms. Results showed a dissociation in false alarm rates, such that in short-term tests there were more false alarms in phonological than sematic lists, whereas in long-term tests there were more false alarms for semantic than phonological lists. Successfully replicating the findings of the previous 6-item study, with the current 4-item study suggests that the results were not due to exceeding working memory capacity. At the short-term, phonological coding appears to drive errors and increases false alarms for phonological lists compared to semantic lists. After a delay the effect reverses, and semantically driven errors increase. Overall, the results suggest distinct processes are involved in the production of false memories in STM compared to LTM.
Marsh, Elizabeth, "Phonological And Semantic False Memories Across Memory Systems" (2021). Psychology. 17.