Evidence for an evaluative effect of stimulus co-occurrence may be inflated by evaluative differences between assimilative and contrastive relations.

Abstract

Recent research on relational evaluative conditioning (relational EC) suggests that stimulus co-occurrence can have a direct effect on evaluations over and above the particular relation between the co-occurring stimuli. This research is based on a process dissociation approach where co-occurrence effects are demonstrated via attenuated evaluative learning for co-occurring stimuli that are connected by contrastive in comparison to assimilative relations. Instead of attributing such attenuations to an orthogonal influence of stimulus co-occurrence, we investigated whether (a) contrastive relations tend to produce weaker evaluations than their assimilative counterparts and (b) such evaluative differences can inflate evidence for co-occurrence effects on continuous as well as on categorical evaluation measures. A pilot study (N = 85) confirmed notion (a), while a first experiment (N = 42) produced preliminary evidence for notion (b) in the context of multinomial processing tree (MPT) modelling. In a second, high-powered experiment (N = 229), sub-sample MPT analyses (including only CSs with correct memory for the CS-US proposition) demonstrated that evidence for co-occurrence effects can be inflated by evaluative differences between assimilative vs. contrastive relations. The theoretical and methodological implications of these findings are discussed.