The intervening task method: implications for measuring mediation.

To study mediation, investigators sometimes examine the effect of an independent variable on an unrelated filler task that precedes the focal task. This approach assumes that the same psychological process drives performance on both tasks. The authors tested this assumption in a stereotype threat paradigm by manipulating whether or not the intervening task was described as relevant to the gender-math stereotype. When performance was relevant to the stereotype, females outperformed controls on an intervening Stroop task, but not when it was irrelevant (Experiment 1). In fact, females anticipating taking a math test under threat withdrew effort and performed more poorly on the intervening task when performance was irrelevant (Experiment 2). These findings suggest that different processes may drive performance on irrelevant and relevant intervening tasks. As a result, performance on irrelevant filler tasks may actually tell scholars little about mediating mechanisms.