Choking and Performance 1 Running Head: CHOKING AND EXCELLING Choking and Excelling Under Pressure

Decrements in performance on cognitive tasks resulting from pressure to perform (i.e., choking) are thought to be caused by interference with the ability to use explicit strategies (the distraction theory). This view suggests that pressure should improve tasks for which explicit strategies hamper performance. This hypothesis was tested by giving participants one of two nearly identical learning tasks, one that required learning a rule, or one that required a holistic information integration strategy. Explicit rule use hurts performance in the latter task. Consistent with the distraction theory, pressure decreases performance on the rule-based task but enhances performance on the information integration task. Choking and Performance 3 Psychological research provides evidence for the anecdotal phenomenon that pressure causes decrements in performance on cognitive and motor tasks (e.g., Beilock & Carr, 2005; Beilock, Kulp, Holt, & Carr, 2004; Masters, 1992). This phenomenon is often called choking under pressure, because people become unable to perform a task that they would otherwise perform well. A prominent explanation for choking under pressure in cognitive tasks is the distraction hypothesis (Beilock & Carr, 2005; Beilock et al., 2004). On this view, pressure leads to a decrease in available working memory resources, which in turn has a negative influence on cognitive performance. In support of this view, for example, Beilock and Carr (2005) found that individuals with high-working memory capacity were more strongly affected by a pressure manipulation than were those with a low workingmemory capacity. An interesting implication of this distraction hypothesis is that pressure might enhance performance on tasks for which explicit working-memory intensive processes harm performance. As an example, Figure 1 shows the stimulus structure for two simple categorization tasks. For each task, the stimuli can be described by two dimensions (e.g., the width and orientation of sine-wave gradients). Figure 1a depicts a simple rule-based task in which the participant must focus on one of the two dimensions and determine the location on that dimension that separates the two categories. This task is typically thought to involve explicit hypothesis testing, and so it should be harmed by any procedure that decreases working memory capacity (Ashby, Alfonso-Reese, Turken, & Waldron, 1998; Maddox & Ashby, 2004; Maddox, Filoteo, Hejl, & Ing, 2004). Consistent with this proposal, Zeithamova and Maddox (in press; see also Maddox, Ashby, Ing & Pickering, 2004; Waldron & Ashby, 2001) showed that learning of this particular category structure was impaired in a dual-task setting relative to a control condition in which no dual task was present. In contrast, the stimulus structure in Figure 1b rotates the category boundary 45degrees in stimulus space. Thus, the rule that separates the categories cannot be stated easily. This stimulus configuration is called an information integration structure, and is thought to be best learned by a procedural or similarity-based process that is limited in its demands on working memory (Maddox & Ashby, 2004; Maddox, Ashby, & Bohil, 2003). Explicit hypothesis testing strategies can be used to solve these tasks, but they lead to suboptimal performance. Even so, previous research suggests that early in learning people start by using hypothesis-testing strategies even in information-integration tasks (Ashby, Alfonso-Reese, Turken & Waldron, 1998). Thus, if people’s working memory capacity were disrupted by pressure, they might abandon hypothesis-testing strategies more quickly and thus improve their ability to acquire this information integration structure. That is, we should observe people excelling under pressure in this condition. We tested this hypothesis in the present study. The manipulation of pressure was similar to that used by Beilock and Carr (2005). Participants in the Low Pressure condition were told that they were going to perform a category learning task and were told to do their best. Participants in the High Pressure condition were told that they and a partner would each receive a performance bonus of $6 in the study if they both exceeded a performance criterion, but that neither would receive the bonus if one of them failed to exceed the criterion. Furthermore, they were told that their partner had already exceeded the criterion, and so the bonus (for they and their partner) was contingent on the Choking and Performance 4 participant’s performance. Beilock and Carr (2005) found that the High Pressure group performed reliably worse than the Low Pressure group on a mathematics task. The distraction view predicts that, relative to the Low Pressure condition, participants in the High Pressure condition should have a more difficult time learning the rule-based categories shown in Figure 1a, but an easier time learning the information integration categories shown in Figure 1b.