Using the process dissociation procedure to estimate recollection and familiarity in working memory: An experimental and individual differences investigation

Although complex span tasks are believed to require controlled attention, many studies have indicated the effects of automaticity on working memory (WM) measures. In two experiments, we disentangled recollection and familiarity using an adaptation of the process dissociation procedure). In two blocks of reasoning span (Experiment 1) and Brown–Peterson span (Experiment 2) tasks, participants studied five discrete to-be-remembered digits (0–9) for either 1 or 3 sec while also solving reasoning problems (Does B follow C? BC). Participants were instructed to report the five presented digits (inclusion) or the five digits that were not presented (exclusion). Estimates of recollection but not familiarity differed as a function of presentation rate. Furthermore, recollection was more reliable than the typically used inclusion performance and strongly related to fluid intelligence. These results help to elucidate the functionally distinct roles of recollection and familiarity in WM.