A problem in measuring the Stroop facilitation and interference effects: implications for measuring performance change in general.

The Stroop facilitation and interference effects as observed in the Stroop color-naming task are typically measured as the difference in naming time between the congruent or the conflicting condition and the neutral condition. The measure assumes the regression function relating the congruent or conflicting naming time to neutral naming time is linear with a slope of one and a nonzero intercept. Instead, analysis showed that the regression function subserving the facilitation effect, in fact, had a slope of less than one with a zero intercept. Analysis also showed that the regression function subserving the interference effect had a zero intercept with a greater than one slope. These results suggest that the Stroop facilitation and interference effects are the consequence of a multiplicative process by which each stage of the naming process is reduced or increased by a constant proportion. Also the appropriate measure of the Stroop effect is the ratio of the congruent (for facilitation) or the conflicting (for interference) naming time over the neutral naming time.