Using latent class modeling to detect bimodality in spacing effect data

A recently proposed theory of the spacing effect [Raaijmakers, J. G. W. (2003). Spacing and repetition effects in human memory: application of the SAM model. Cognitive Science, 27, 431–452.] suggests that the spacing effect is conditional on study-phase retrieval leading to two groups of students showing different magnitudes of the spacing effect. This bimodality was also observed in histograms of spacing-effect data. In this study, we used latent class regression analysis to investigate whether these groups can be detected in existing datasets (Experiment 1). Specific hypotheses about the magnitude of the spacing effect in the latent classes were assessed in Experiment 2. Latent class regression analysis in both experiments showed that the fit of the two-class model was considerably better than the (1-class) ANOVA model. Moreover, the results of Experiment 2 showed, in line with our predictions, that when the presentation rate changed from 1 s to 4 s the increase in spacing effect was larger for the low-performing class than for the high-performing class.