| Last Updated:
Creating DOI. Please wait...
Evaluative statements (ES; “X will be associated with pleasant and Y with unpleasant”) have been shown to result in stronger shifts in implicit (IAT) attitudes than repeated evaluative pairings (REP) of X and Y with pleasant and unpleasant images. Experiment 1 (N = 4217) demonstrates that whereas implicit attitudes produced via REP remain relatively stable over time, the overall advantage of ES over REP is driven exclusively by early trials, followed by rapid deterioration even within a single IAT session. Experiment 2 (N = 2365) provides evidence that attitude shifts created via REP are dependent on contingency awareness but are not modulated by the number of stimulus pairings presented. Overall, these results suggest that REP shifts implicit attitudes by giving rise to propositions about the valence of attitude objects, with such propositions being initially less strong but more durable than ones based exclusively on verbal information.