What makes working memory work? A multifaceted account of the predictive power of working memory capacity

Working memory capacity (WMC) has received a great deal of attention in cognitive psychology partly because WMC correlates broadly with other abilities (e.g., reading comprehension, second-language proficiency, fluid intelligence), and thus seems to be a critical aspect of cognitive ability. However, it is still rigorously debated why such correlations occur. Some theories posit a single ability (e.g., attention control, short-term memory capacity, controlled memory search) as the primary reason behind WMC’s predictiveness, whereas others argue that WMC is predictive because it taps into multiple abilities. Here, we tested these single- and multifaceted accounts of WMC with a large-scale (N = 974) individual-differences investigation of WMC and three hypothesized mediators: attention control, primary memory, and secondary memory. We found evidence for a multifaceted account, such that no single ability could fully mediate the relation between WMC and higher-order cognition (i.e., reading comprehension and fluid intelligence). Further, such an effect held regardless of whether WMC was measured via complex span or n-back.