Jacoby et al. (Psychonomic Bulletin & Review, 10, 638–644, 2003) reported that, in Stroop tasks, stimuli that more frequently involve targets combined with a congruent distractor (e.g., the word RED in the color red) produce larger Stroop effects than stimuli that more frequently involve targets combined with an incongruent distractor (e.g., RED in green). This pattern suggests that adaptive control can be item-specific in addition to item-nonspecific, and reactive in addition to proactive (although this conclusion has been challenged). This adaptive-control process has often been assumed to be driven by the conflict associated with incongruent stimuli; however, the typical experimental manipulations investigating this issue allow the facilitation associated with congruent stimuli to also play a role. Here, we modified those manipulations in order to focus exclusively on conflict, removing any impact of congruency facilitation, by contrasting targets presented with either neutral (letter strings) or incongruent distractors. Neutral stimuli were presented more frequently than incongruent ones in the Mostly-Neutral (MN) condition and vice versa in the Mostly-Incongruent (MI) condition. Paralleling the original pattern, Stroop interference was larger in the MN condition, suggesting that item-specific conflict frequency can be used to adapt attention accordingly. Importantly, this effect was replicated after experimentally controlling for stimulus frequency, a confound that was found to explain part, but not all, of the general pattern. These results support Jacoby et al.’s claims that (a) control can be adapted in an item-specific fashion and (b) conflict plays a key role in that process.
