Evaluative conditioning is sensitive to the encoding of CS-US contingencies

Evaluative conditioning (EC), the change in the evaluation of a neutral "conditioned" stimulus (CS) that is due to its pairing with a liked or disliked "unconditioned" stimulus (US; De Houwer, 2007), has only sometimes been found to depend on the CS-US contingency, that is, on the predictive relation between CS and US occurrences. In the research reported here, we hypothesize that this is a result of the way EC procedures can bias the encoding of CS-US contingencies and therefore bias contingency learning. This may have prevented previous investigations from detecting contingency effects despite EC being sensitive to the encoding of CS-US contingencies. In support of this hypothesis, we show that measured (Experiments 1 a and 1 b) and manipulated (Experiment 2) differences in contingency learning predict the effect of CS-US contingency on EC. Implications for the underlying processes of EC are discussed.