The effect of crowding arrows on Spatial Stroop task: reversing the congruency effect

Recent spatial Stroop studies have reported a reversed congruency effect (RCE) with gaze stimuli—faster responses on incongruent than congruent trials—contrasting with the standard congruency effect (SCE) typically observed with arrows. This difference also emerges in accuracy at faster response latencies (i.e., higher accuracy for incongruent than congruent trials), as shown by conditional accuracy function analyses. Although social significance has been proposed to explain the RCE for gaze, alternative accounts highlight non-social factors, such as the difficulty of extracting task-relevant information. In two separate experiments, with a priori preregistration of hypotheses, sample size, and plan of analyses, we examined whether increasing the difficulty of extracting task-relevant information from arrow stimuli could induce an RCE analogous to that found with gaze. We introduced overlapping bars at three crowding levels. In Experiment 1, we hypothesized that high crowding (i.e., level 3) would reduce or eliminate the SCE by facilitating the decay or active suppression of irrelevant spatial codes. Contrary to our expectation that crowding would merely weaken rather than reverse the SCE, arrows at the higher difficulty levels (2 and 3) consistently produced an RCE. The pattern of results was replicated in Experiment 2, now confirming the a-priori hypothesis that a RCE would be observed with crowding. Notably, not only did these conditions reverse the reaction time patterns, but also eliminated the conflict effect on accuracy in Experiment 1 and showed higher accuracy for incongruent than congruent trials within the fastest responses in Experiment 2. These findings, which mirror the pattern previously observed with gaze stimuli, challenge the view that the RCE arises exclusively from social processing. Instead, they suggest that the reversion of the spatial congruency effect may be due to the masking of relevant information.