Detrimental Influence of Contextual Change on Spacing Effects in Free Recall Peter P. J. L. Verkoeijen, Remy M. J. P. Rikers, and Henk G. Schmidt Erasmus University Rotterdam Two experiments were conducted to determine the mechanism underlying the spacing effect in free-recall tasks. Participants were required to study a list containing once-presented words as well as massed and spaced repetitions. In both experiments, presentation background at repetition was manipulated. The results of Experiment 1 demonstrated that free recall was higher for massed items repeated in a different context than for massed items repeated in the same context, whereas free recall for spaced items was higher when repeated in the same context. Furthermore, a spacing effect was shown for words repeated in the same context, whereas an attenuated spacing effect was revealed for words repeated in a different context. These findings were replicated in Experiment 2 under a different presentation background manipulation. Both experiments seem to be most consistent with a model that combines the contextual variability and the study-phase retrieval mechanism to account for the spacing effect in free-recall tasks. The term spacing effect refers to the phenomenon that repeated items induce better recollection if both occurrences are separated by time or other targets (i.e., spaced presentation), compared with a situation in which repetitions occur in immediate succession (i.e., massed presentation). Although the spacing effect has been dem- onstrated in a vast number of studies (e.g., Challis, 1993; Greene, 1989, 1990; Greene & Stillwell, 1995; Hintzman & Block, 1973; Hintzman, Summers, & Block, 1975; Mammarella, Russo, & Avons, 2002; Russo & Mammarella, 2002; Russo, Mammarella, & Avons, 2002; Russo, Parkin, Taylor, & Wilks, 1998) it has largely defied a unitary explanation (for reviews on proposed theoretical mechanisms see Crowder, 1976; Dempster, 1996; Hintzman, 1974, 1976; Kintsch, 1970). For instance, different mechanisms are proposed to account for spacing effects in cued memory tasks and in free recall. Regarding the spacing effect in free recall, two important the- oretical explanations can be distinguished. The contextual vari- ability account (e.g., Melton, 1967, 1970) suggests that the number of encoded retrieval cues increases with repetition spacing and that the chance that an item is recalled is positively related to the number of retrieval cues. Evidence for the contextual variability account has been obtained in a number of studies using different operationalizations of contextual variability (e.g., Delarosa & Bourne, 1985; Durgunoglu & Roediger, 1987; Glenberg, 1979; Krug, Davis, & Glover, 1990; Madigan, 1969). Alternatively, the study-phase retrieval account states that the storage of contextual features in a repeated item’s memory trace will only take place if a prior presentation is retrieved from long-term store during study (Greene, 1989). This account predicts that spacing will be positively related to free-recall performance under the condition of successful study-phase retrieval: For items that are not retrieved at their second occurrence, the spacing effect will disappear. Moreover, according to the study-phase retrieval point of view, context variation is expected to have no effect on free recall of massed repetitions, because the first presentation is expected still to be in the short-term buffer at its second occur- rence, and therefore it is not retrieved from long-term store. Empirical support for the study-phase retrieval account has been demonstrated in several studies (e.g., Braun & Rubin, 1998; Johnston & Uhl, 1976; Thios & D’Agostino, 1976; Toppino & Bloom, 2002; Toppino, Hara, & Hackman, 2002). Although in the literature the spacing effect in free recall is attributed to either a contextual variability or a study-phase re- trieval mechanism, it might be possible that both mechanisms are required to explain the spacing effect in free-recall tasks. It should be noted that such a combined model predicts different effects of context manipulation for massed and spaced repetitions. On the basis of the contextual variability component of the model, it can be inferred that repetitions in different environmental contexts should benefit massed items more than spaced items, relative to same context repetitions, and consequently recall of massed items should be greater when repeated in a different context. However, on the basis of the study-phase retrieval component of such model, the reversed pattern is expected for spaced items. The encoding specificity principle states that the probability of retriev- ing a particular memory event is positively related to the degree of overlap between information in the retrieval cue and the context information stored in the event’s memory trace (Tulving, 1983; Tulving & Thomson, 1973; see also Smith & Vela, 2001, for information on context-dependent memory). An important impli- cation of the encoding specificity principle is that repetitions in different contexts should impair performance in the spaced condi- tion, because prior occurrences in a different context will be relatively difficult to retrieve. This implication has been repeatedly Peter P. J. L. Verkoeijen, Remy M. J. P. Rikers, and Henk G. Schmidt, Department of Psychology, Erasmus University Rotterdam, Rotterdam, the Netherlands. We thank Kenneth J. Malmberg for valuable comments on drafts of this article and Manon Augustus for her assistance in collecting the data. Correspondence concerning this article should be addressed to Peter P. J. L. Verkoeijen, Department of Psychology, Erasmus University Rot- terdam, P. O. Box 1738, NL-3000 DR, Rotterdam, the Netherlands. E-mail: verkoeijen@fsw.eur.nl Journal of Experimental Psychology: Copyright 2004 by the American Psychological Association Learning, Memory, and Cognition 2004, Vol. 30, No. 4, 796 – 800 0278-7393/04/$12.00 DOI: 10.1037/0278-7393.30.4.796 796