Measuring the Effect of Display Frequency

• Neutral trials were used to measure the display frequency effect • All features had equal marginal frequency • The true proportion congruency effect was estimated by subtracting neutral trials from the congruency trials Proportion Congruency Effect • The Stroop effect is known to be dependent on the proportion of congruent trials: larger when there are more congruent trials in a block (e.g., Tzelgov, Henik, & Berger, 1992) • This had been shown to be true for a within block manipulation: words which are mostly congruent (e.g., GREEN and BLUE) show larger Stroop effect than words which are mostly incongruent (e.g., WHITE and YELLOW; Jacoby, Lindsay & Hessels, 2003) • The manipulation of congruency was confounded with the display frequency of the individual item pairs (Schmidt & Besner, 2008) • The congruent trials in the low proportion congruent condition and the incongruent trials in the high proportion congruent condition were more frequent, suggesting an alternative display frequency accoun