Spacing and repetition effects in human memory: application of the SAM model

Abstract

Spacing between study trials of an item increases the probability that item will be recalled. This article presents a new model for spacing based on the SAM theory of memory developed by Raaijmakers and Shiffrin (1980, 1981). The model is a generalization of the SAM model as applied to interference paradigms (Mensink & Raaijmakers, 1988, 1989 and may be viewed as a mathematical version of the Component-Levels theory proposed by Glenberg (1979). It is assumed that on a second presentation of an item, information is added to an existing trace if the episodic memory image corresponding to that item is retrieved. If it is not retrieved, a new image is stored. It is shown that the model predicts many standard findings including the lack of a spacing effect for the recall of at least one of two items each presented once (Ross & Landauer, 1978).

Electrophysiological correlates of word repetition spacing: ERP and induced band power old/new effects with massed and spaced repetitions

“…The spacing effect on free recall has been explained as a consequence of either deficient processing during massed repetitions (e.g., Challis, 1993) or the storage of additional retrieval cues in the memory trace during spaced repetitions (Raaijmakers, 2003;Verkoeijen et al, 2004). The present ERP word repetition effects suggest that massed repetition, when compared to spaced repetition, is accompanied by more automatic and less controlled processing than spaced repetition as indexed by the N400 and LPC repetition effects.…”

“…It should be noted that the alpha 'old/new' effect mainly takes place after the participant has give a correct 'old' response (about 760 ms after stimulus onset, see Table 1), that is, after correct retrieval. It could be that this late cortical activation is related to the more elaborate encoding (i.e., encoding of additional cues) that characterizes spaced repetitions (Raaijmakers, 2003), although the correlational analysis did not reveal any association between the upper alpha repetition effect and subsequent performance on the free recall task.…”

“…In the latter case, the repeated item will seem very familiar and the participant will erroneously think that the item is already well learned (Greene, 1989). Alternatively, the spacing effect has been explained as a consequence of the storage of additional retrieval cues in the memory trace of the repeated item (Raaijmakers, 2003;Verkoeijen et al, 2004). With spaced repetition the item is retrieved from long-term store (LTS) at the second presentation (if recognized) and additional information (contextual elements such as the temporal lag between the first and second presentation) will be added to the original LTS memory trace.…”

Executive control of retrieval in noun and verb generation

“…Although the present paper purposely focused on noun/verb generation, extending the model to other memory paradigms may reveal deep commonalities in control and associative mechanisms involved in different tasks, and thus advance our knowledge of general memory principles (Surprenant & Neath, 2009). In this regard, it is interesting to note that, being based on a general model of associative retrieval, our proposed account bears clear resemblances with retrieve/test process models of recall (like the Search of Associative Memory: Raaijmakers & Shiffrin, 1981; see also Raaijmakers, 2003), which posit an associative process of sampling candidates for recall followed by a subsequent editing/recognition process.…”

Using Multidimensional Encoding and Retrieval Contexts to Enhance Our Understanding of Stochastic Dependence in Source Memory

“…We designed this figure to represent the likelihood that each item presentation potentially creates a separate episodic representation (cf. Raaijmakers, 2003). Their reasoning was that if the item is a mediator between context information in the memory trace, then as long as each context was established strongly to the item, it should not matter whether these respective associations were encoded contemporaneously or separately.…”