In the affective misattribution procedure (AMP), pairs of prime and target stimuli appear rapidly in succession. Attitudes toward the prime influence the evaluation of the target despite instructions to avoid this influence. Because this priming effect presumably happens without people’s knowledge, the AMP is used to study automatic evaluation. Participants in four studies performed the AMP and reported their perception of the priming effect. The authors found that the priming reflected reliable and valid attitudes toward the primes mostly among participants who reported that the priming occurred and that they intentionally rated the primes instead of the targets. The authors conclude that the AMP hardly captures attitude effects that escape people’s knowledge. The AMP’s good psychometric qualities as an attitude measure rely mainly on a small subset of participants who believe that they intentionally caused the attitude effect.