Examining the Episodic Context Account of the Testing Effect: Does episodic recall enhance memory for context?

Abstract

The benefits of testing or retrieval practice on subsequent memory are firmly established, yet to date, the underlying mechanism remains unclear. According to the Episodic Context Account by Karpicke, Lehman, and Aue (2014), episodic recall enhances memory by reinstating and strengthening the episodic context associated with the retrieved memory. However, this account remains ambiguous as to which types of contexts are part of the retrieved and strengthened context. We aimed to refine this account by establishing which kinds of source information are strengthened following retrieval practice. Through Experiments 1 – 4, we provide direct evidence that memory for item specific source features (i.e., font colors) are not strengthened following retrieval practice. In Experiments 5 and 6, we test a possibility that only source characteristics known to guide memory organization are strengthened following retrieval practice (temporal: Kahana, 1996; Healey, Long, & Kahana, 2018; spatial: Miller et al., 2013a, b). Retrieved-context models propose that the contiguity effect, a tendency for neighboring study items to be recalled successively, is a marker of reinstated temporal context. If retrieval practice of an item reinstates the temporal context associated with the item’s local neighbors on the list, then the benefits of retrieval practice should spread to these neighbors as well. Similarly, a wealth of studies demonstrate that semantic category exerts powerful organizational effects during memory search (Bousfield, 1953). It may be that the benefits of retrieval practice spread to semantic associates of the practiced items. We examine this question through characterizing the effects of retrieval practice on an item’s original neighbors and semantic associates from the study list. The present results suggest that context reinstatement as a mechanism for the retrieval practice benefit may be contingent on several factors such as the context feature’s relevancy to the retrieved item or saliency as a recall cue.