Of Two Minds: A registered replication

Several dual-process theories of social evaluation posit that implicit (or automatic) and explicit (or non-automatic) evaluations reflect distinct attitudes acquired via qualitatively different learning processes. Consequently, one may explicitly like a person but dislike them implicitly. Rydell et al. (2006) reported a striking demonstration of this dissociation: Briefly flashed negative words paired with a person described as behaving positively produced positive ratings of that person but negative Implicit Association Test scores. However, Heycke et al. (2018) could not replicate this dissociation. In two new replication attempts (N = 593) across four countries and three languages, with words flashed between 13 and 27 ms, we too found overwhelming evidence against the dissociation between directly and indirectly measured evaluations. We conclude that the dissociative evaluative learning effect reported by Rydell et al. (2006) is replicably non-replicable.