Training on integrated versus separated Stroop tasks: the progression of interference and facilitation.

Two experiments examined the course of interference and facilitation in the Stroop (1935) task during training. Two versions of the task were compared: integrated (e.g., the word RED in the color green) and separated (e.g., green asterisks above the word RED). Stimuli were congruent (RED in red), incongruent (GREEN in red), or neutral (XXX in red). Over 5-day (Experiment 1) and 10-day (Experiment 2) training sessions, facilitation due to congruence was small, stable, and equivalent in both task versions. In contrast, interference declined sharply on the integrated task over Days 1-3, then slowed to parallel the gradual decline on the separated task. Finally, training on the color naming task did not affect a word reading task administered after training. These findings imply that (a) Stroop interference initially reflects two problems--overcoming integration and managing two conflicting information sources; (b) with practice, the larger integration problem is solved relatively quickly, rendering the integrated and separated tasks quite comparable thereafter; and (c) facilitation and interference in the Stroop task may be independent. These results challenge extant theories of the Stroop effect, which do not predict such effects.