Effect of circadian rhythms on retrieval-induced forgetting

We paid particular attention in selecting anagrams whose solutions were far from interfering with the experimental categories and items. None of the solutions of the anagrams were close to creating interference with the recall of experimental categories and items.

Before breaking the results down by two separate 2 × 2 ANOVAs, we conducted a 2 × 2 × 2 mixed-model ANOVA with item strength (strong vs. weak items) and item type (Nrp and Rp− items) as repeated measures factors and synchrony (optimal times vs. non-optimal times) as a between subject factor. This analysis revealed a significant main effect of item strength, F (1, 50) = 11.80, p < .001, pη ² = .19, and a significant main effect of item type, F (1, 50) = 11.04, p < .001, pη ² = .18. Furthermore, the interaction between item type and synchrony and the interaction between item type and item strength were both significant, F (1, 50) = 7.52, p < .01, pη ² = .13 and F (1, 50) = 10.15, p < .01, pη ² = .17, respectively. Finally, the three-way interaction was not significant, F (1, 50) = 1.32, p > .2, pη ² = .03. Possibly, this last result is due to a general lack of statistical power. However, in order to better address whether synchrony differently affected RIF for strong and weak categories, we conducted two separate ANOVAs, one for high taxonomic frequency and one for low taxonomic frequency exemplars (see Anderson et al. 1994, for similar procedure).