Unmasking the component-general and component-specific aspects of primary and secondary memory in the immediate free...

The immediate free recall (IFR) task has been commonly used to estimate the capacities of the primary memory (PM) and secondary memory (SM) components of working memory (WM). Using this method, the correlation between estimates of the PM and SM components has hovered around zero, suggesting that PM and SM represent fully distinct and dissociable components of WM. However, this conclusion has conflicted with more recent studies that have observed moderately strong, positive correlations between PM and SM when separate attention and retrieval tasks are used to estimate these capacities, suggesting that PM and SM represent at least some related capacities. The present study attempted to resolve this empirical discrepancy by investigating the extent to which the relation between estimates of PM and SM might be suppressed by a third variable that operates during the recall portion of the IFR task. This third variable was termed "strength of recency" (SOR) in the present study as it reflected differences in the extent to which individuals used the same experimentally-induced recency recall initiation strategy. As predicted, the present findings showed that the positive correlation between estimates of PM and SM grew from small to medium when the indirect effect of SOR was controlled across two separate sets of studies. This finding is important because it provides stronger support for the distinction between "component-general" and "component-specific" aspects of PM and SM; furthermore, a proof is presented that demonstrates a limitation of using regression techniques to differentiate general and specific aspects of these components.