g., Unsworth, Redick, et al., 2009). WM processing was also weakly and negatively correlated with

capacity and SM. However, WM processing demonstrated a moderate correlation with AC suggesting that AC abilities are needed during the processing components of complex span tasks. Thus, WM processing and WM storage demonstrated differential relations with capacity, AC, and SM, with WM storage being moderately related to all, but WM processing being more related to AC than to capacity or SM. Our next model examined whether WM processing would account for the relation between WM storage and gF or whether both would contribute independently selleck to gF. To examine this we specified a model in which both WM storage and WM processing predicted gF and WM storage and WM processing were correlated. If WM processing accounts for the relation between

WM storage and gF we should see that WM processing and WM storage are related, but only WM processing significantly predicts gF. If both contribute independently to gF we should see that both predict gF. The fit of the model was acceptable (see Table 3). As shown in Fig. 5 both WM storage and WM processing predicted gF. Collectively, 50% of the variance in gF was accounted for with WM storage uniquely accounting for 18%, WM processing uniquely accounting for 21%, and both shared 11% of the variance. Consistent with prior research these results suggests that WM storage and WM processing make independent contributions to higher-order cognition and in particular to gF (Bayliss et al., 2003, Logie and Duff, 2007, Unsworth et al., 2009 and Waters and Caplan, selleck inhibitor 1996). For our final model we examined whether capacity, AC, and SM would

mediate the relations between WM storage and WM processing with gF. That is, similar to the model shown in Fig. 3, we wanted to examine whether capacity, AC, and SM would mediate not only the relation between WM storage and gF, but also the relation between WM processing and gF. Therefore, we specified a model in which WM storage and WM processing were correlated and both predicted capacity, AC, and SM. The paths from WM storage and WM processing to gF were set to zero. Capacity, attention control, and secondary memory, were specified to predict gF. As shown in Table Reverse transcriptase 3 the fit of the model was good. Shown in Fig. 6 is the resulting model. As can be seen, WM storage was significantly related to capacity, AC, and SM. Likewise, WM processing was related to capacity, AC, and SM, but the strongest relation was with AC. Furthermore, capacity, AC, and SM all significantly predicted gF with 81% of the variance being accounted for in gF. Importantly, freeing the paths from WM storage and WM processing to gF did not change the model fit (Δχ2(2) = 3.98, p > .14), indicating that the paths were not significant and did not uniquely predict gF.