Distinguishing Response Conflict and Task Conflict in the Stroop Task: Evidence From Ex-Gaussian Distribution Analysis Marco Steinhauser and Ronald Hu ¨bner Universita ¨t Konstanz, Konstanz, Germany It has been suggested that performance in the Stroop task is influenced by response conflict as well as task conflict. The present study investigated the idea that both conflict types can be isolated by applying ex-Gaussian distribution analysis which decomposes response time into a Gaussian and an exponential component. Two experiments were conducted in which manual versions of a standard Stroop task (Experiment 1) and a separated Stroop task (Experiment 2) were performed under task-switching conditions. Effects of response congruency and stimulus bivalency were used to measure response conflict and task conflict, respectively. Ex-Gaussian analysis revealed that response conflict was mainly observed in the Gaussian component, whereas task conflict was stronger in the exponential component. Moreover, task conflict in the exponential component was selectively enhanced under task-switching conditions. The results suggest that ex-Gaussian analysis can be used as a tool to isolate different conflict types in the Stroop task. Keywords: Stroop task, task switching, response time distribution, cognitive conflict The Stroop task (Stroop, 1935) is one of the most frequently applied paradigms in cognitive psychology (MacLeod, 1991). It requires that participants name a word’s color but ignore its meaning, which refers to a different color. The Stroop effect denotes the finding that the latency and accuracy of color naming is strongly affected by the word’s meaning. For instance, naming the font color of a word shown in blue takes longer and is more prone to error when the meaning of the word is “red” than when the meaning is “blue.” The Stroop task often serves as a tool for investigating cognitive processes such as selective attention (e.g., Spieler, Balota, & Faust, 2000), automaticity (e.g., Cohen, Dunbar, & McClelland, 1990), and reading (e.g., Masson, Bub, Woodward, & Chan, 2003). In recent years, the Stroop task was also used to examine mechanisms involved in cognitive control and conflict resolution (e.g., Botvinick, Braver, Barch, Carter, & Cohen, 2001). These studies typically assumed that the Stroop effect results from a conflict between relevant and irrelevant representations. However, despite the huge body of research on this issue, relatively little work has been conducted to reveal the level(s) on which these conflicts occur. The goal of the present study was to deter- mine the contribution of different conflict types to Stroop perfor- mance. Using ex-Gaussian distribution analysis (Heathcote, Pop- iel, & Mewhort, 1991), we provide evidence that task conflicts and response conflicts are reflected in different features of the response time (RT) distribution. Response Conflict and Task Conflict The Stroop effect is frequently explained in terms of a response conflict (e.g., Cohen et al., 1990; Hunt & Lansman, 1986; Roelofs, 2003). In this view, the word automatically activates the phono- logical code of its meaning. If this code does not correspond to the phonological code of the color, a response conflict emerges that delays responding. In addition, Monsell, Taylor, and Murphy (2001) proposed that performance in the Stroop task is also influ- enced by a task conflict. They assumed that a Stroop stimulus activates not only two competing responses but also two compet- ing tasks: color naming and word naming. To corroborate this idea, Monsell et al. (2001) showed that the interfering influence of neutral words on color naming is independent of word frequency, which suggests that the effect is mediated by stimulus–task asso- ciations. However, this result provides only indirect evidence for the idea of a task conflict. Moreover, it has been contradicted by recent evidence showing that word frequency can influence color naming under some conditions (Burt, 2002). For a closer examination of the effects of different conflict types in the Stroop paradigm, it would be helpful to have a direct indicator not only of the strength of response conflict but also of the strength of task conflict. In recent years, such indicators have been examined in the task-switching paradigm, in which partici- pants alternate between different tasks (Allport, Styles, & Hsieh, 1994; Meiran, 1996; Rogers & Monsell, 1995). Task-switching studies often focus on the switch costs, that is, the performance decrement on task-switching trials relative to task-repetition trials. However, the paradigm can also be used to examine the influence of stimulus-induced conflicts. For instance, Rogers and Monsell (1995) as well as Steinhauser and Hu ¨bner (2007) used stimuli that contained a target associated Marco Steinhauser and Ronald Hu ¨bner, Fachbereich Psychologie, Uni- versita ¨t Konstanz, Konstanz, Germany. This research was supported by a grant to Ronald Hu ¨bner from the Deutsche Forschungsgemeinschaft (DFG: Hu 432/8). We thank Andrew Heathcote, Michael Masson, and an anonymous reviewer for helpful com- ments. Correspondence concerning this article should be addressed to Marco Steinhauser, Fachbereich Psychologie, Universita ¨t Konstanz, Fach D29, D-78457 Konstanz, Germany. E-mail: Marco.Steinhauser@uni-konstanz.de Journal of Experimental Psychology: © 2009 American Psychological Association Human Perception and Performance 2009, Vol. 35, No. 5, 1398 –1412 0096-1523/09/$12.00 DOI: 10.1037/a0016467 1398