Previous research suggests that evaluative conditioning(EC) is insensitive to contingency awareness, andinsensitive to extinction and contingency manipulations.Recent studies have tested the generality of these ECresults using physiological (startle reflex) andbehavioural (affective priming) measures of affect ratherthan verbal ratings of liking. The current experimentsused an affective priming measure to extend this programof work. Over four experiments we examined the role ofawareness, the role of contingency, and the phenomena ofcue-competition in evaluative conditioning. Eachexperiment involved subjects experiencing conditionedstimuli (CSs) paired with affectively valenced (positiveand negative) visual images. Subjects then classifiedwords as either positive or negative. Each word waspreceded by presentation of a CS. It was found thatsubjects classified the words faster if the valence ofthe word was the same as that of the immediatelypreceding CS. However, this conditioning effect did notdepend on contingency awareness, was not reduced bycontingency degradation, and there was no evidence ofcue-competition in a blocking procedure that used anovershadowing control.