Comparison of semantic and phonological false memories in short- and long-term tests.

This article investigates how false memories differ for phonologically and semantically related word lists across short-term memory (STM) and long-term memory (LTM) delays using the Deese–Roediger–McDermott (DRM) paradigm. Across three experiments with undergraduate and online participants, results consistently showed higher false recognition of phonologically related critical items (CIs) immediately after study (STM), whereas semantically related CIs elicited more false alarms after a 1-minute delay (LTM). These findings suggest that STM retrieval relies more on phonological (surface-level) information, while LTM retrieval depends more on semantic (meaning-based) information. The study also examined theoretical accounts—activation-monitoring framework (AMF) and fuzzy-trace theory (FTT)—finding that both explain many results but face challenges fully accounting for the dissociation in false memory patterns across delays. Overall, the research supports the view that semantic and phonological information contribute differently to memory errors depending on retention interval, highlighting the interplay between STM and LTM processes without strictly separating them as distinct systems.