When does an extinction procedure lead to mere exposure effects and extinction of evaluative conditioning?

Repeatedly presenting a neutral stimulus (conditioned stimulus, CS) together with a positive or negative stimulus (unconditioned stimulus, US) typically changes liking of the CS. An important question is whether a subsequent extinction phase where the CS is presented without the US extinguishes such evaluative conditioning (EC) effects. In this regard, it is crucial to consider that an extinction procedure involves repeated exposure to the CS that is known to influence evaluations due to exposure-induced increases in processing fluency. However, it is unclear whether exposure-induced changes in CS liking occur during extinction and hence bias observed extinction effects. In three experiments (NExperiment1 = 301, NExperiment2 = 296, NExperiment3 = 560), with pictures and words as stimuli, we observed that repeated CS-only presentations after conditioning increased CS processing fluency. Moreover, CS-only presentations also increased CS liking, but only when participants had previously evaluated stimulus fluency. When controlling for mere exposure effects on liking, we observed extinction of EC only when participants evaluated the CSs both post-conditioning and post-extinction, replicating previous findings. We outline implications for EC methodology and applied research and discuss the findings in terms of memory-based judgment processes.