Does Evaluative Learning Rely on the Perception of Contingency?

An experiment is reported studying the impact of objective contingency and contingency judgments on cross-modal evaluative conditioning (EC). Both contingency judgments and evaluative responses were measured after a contingency learning task in which previously neutral sounds served as either weak or strong predictors of affective pictures. Experimental manipulations of contingency and US density were shown to affect contingency judgments. Stronger contingencies were perceived with high contingency and with low US density. The contingency learning task also produced a reliable EC effect. The magnitude of this effect was influenced by an interaction of statistical contingency and US density. Furthermore, the magnitude of EC was correlated with the subjective contingency judgments. Taken together, the results imply that propositional knowledge about the CS-US relationship, as reflected in contingency judgments, moderates evaluative learning. The data are discussed with respect to different accounts of EC.