Abstract
The present study modeled the relationship between the Stroop and the neutral naming times to investigate the mechanism underlying the Stroop interference effect. 95 subjects took six alternative versions of the Stroop task and the naming times in the Stroop and neutral conditions were each averaged across the tasks to arrive at a more general measure of the Stroop and neutral naming times. Regression analysis of these general measures indicated a linear function with a small and nonsignificant intercept and a slope significantly greater than one. This finding is consistent with Chen's 1996 results and supports the hypothesis that the mechanism underlying the Stroop interference is interactive or multiplicative rather than stage-like or additive and that a ratio of Stroop over neutral naming times was psychologically a more appropriate measure.