Recency is Sufficient for Reconciling Categorisation and Memory: Commentary on Devraj et al. (2024)

Devraj et al. (Psychonomic Bulletin & Review. doi:10.3758/s13423-023-02448-2, 2024) presented evidence which they claimed demonstrated that people adapt their categorisation strategies to task demands. When items are distributed uniformly across time as is the standard procedure in categorisation experiments, participants initially classify items based on similarity to category prototypes. Over time, as more detailed exemplar representations of categories are formed, people shift towards utilising exemplar-based classification for increased performance. By contrast, if items are experienced less frequently over time as is commonly observed in real-world conditions, participants instead seem to initially employ an exemplar strategy and then shift towards prototype classification. By re-analysing the same data, we argue instead that this pattern can be produced with exemplar-forgetting in both conditions. By systematically increasing in the delays across which stimuli were tested, their Experimental condition exaggerated the effects of forgetting on performance over time compared to the Control condition. This resulted in a reversal of performance growth over time – instead leading to a steady decline in performance. As exemplar model-fit advantage is expected to vary with performance, we suggest that trends in this advantage are not diagnostic of a shift in classification strategy. We found that not only did a forgetting-function improve exemplar model fit to Devraj et al.’s data, under reasonable parameters this model could predict the observed patterns of both performance and model-fit seen in both conditions a-priori. When compared with a strategy-shifting mixture model, exemplar-forgetting provided equivalent fits despite being both parametrically and theoretically more parsimonious.