The Impact of Interference Rate on the Test-Retest Reliability of the Stroop Test

Interference tasks (e.g., the Stroop task) exhibit low test-retest reliability, which hinders the measurement of individual differences. This study examined whether controlling the interference rate improves reliability. The interference rate is a metric derived from the Horse-Race Model, defined as the probability of experiencing semantic-color conflict in incongruent Stroop trials (distinct from list-wide congruency, i.e., the proportion of incongruent trials in a block). In Studies 1a and 1b, interference rate was positively correlated with the magnitude of the Stroop effect, indicating that it accounts for meaningful individual variability. In Study 2, the fixed-interference-rate group (IR-fixed) had their interference rate controlled at 75% across two testing sessions (2-week interval) by adjusting the stimulus-onset asynchrony between semantic and color features (i.e., ensuring a 75% probability of experiencing interference in incongruent trials). Compared to the random-interference-rate group (IR-random), the IR-fixed group showed significantly higher test-retest reliability and stronger cross-session predictability of the Stroop effect, demonstrating that controlling the interference rate enhances measurement stability. Studies 1b and 2 also revealed moderate cross-session correlations for interference rate, regardless of whether the interval was a single Stroop task administration or 2 weeks. Furthermore, across all studies, interference rate exhibited poor internal consistency (McDonald's ω < .60). Overall, the results suggest that interference rate can predict the Stroop effect but is not a stable psychological trait; instead, it fluctuates within a certain range. Therefore, failing to control the interference rate adversely impacts the test-retest reliability of the Stroop task.