Stimulus-Response Mapping Shapes Patterns of Interference in the Arrow-Word Stroop Task

The arrow-word Stroop task measures inhibitory control by assessing the slowing of vocal responses to arrow direction when paired with an incongruent word (e.g., ‘LEFT’ printed on a right-pointing arrow). This interference is commonly attributed to the automaticity of word reading. However, the same pattern is not observed with manual responses, suggesting modality influences interference. We investigated competing theoretical accounts by systematically manipulating the task-relevant stimulus and response modality. Thirty-two participants (18–52 years) completed two Stroop variants using either manual left/right button presses, or vocal “left”/“right” responses. In each condition the task-relevant stimulus was either the arrow or word, and performance in Pure (100% congruent or incongruent trials) and Mixed (66% congruent, 33% incongruent) blocks were compared with Baseline blocks. Vocal responses were faster to words than arrows, and incongruent words slowed responses to arrows, but arrow direction did not affect word responses. This pattern reversed for manual responses: button presses were faster to arrows than words, with incongruent arrows slowing word responses. These findings challenge accounts attributing Stroop interference to automatic word processing, instead suggesting interference arises when the task-relevant stimulus requires translation into the response modality while the irrelevant dimension is more directly compatible with that response.