Does the surveillance paradigm provide evidence for unconscious evaluative conditioning? A Bayesian perspective
In evaluative conditioning (EC), a neutral conditioned stimulus acquires the valence of an intrinsically valenced unconditioned stimulus as a result of repeated pairings between the two. Evidence for EC in the absence of awareness of the pairings has been provided using the surveillance paradigm (Olson & Fazio, 2001). In this commentary, we report a Bayesian reanalysis of data from a multi-lab replication of this paradigm (Moran et al., 2020). Unlike the original frequentist analyses, we conclude that the replication does not provide convincing evidence in favor of unconscious EC under any of the exclusion criteria, including the ones originally used by Olson and Fazio (2001). Moreover, when informative priors relying on the totality of meta-analytic evidence are used, three of four analyses provide evidence against an effect. This commentary highlights the benefits of using Bayesian analytic approaches and cautions against inferring the presence of unconscious EC in the surveillance paradigm.