The construct validity of the affect misattribution procedure (AMP) has been challenged by theories proposing that the task does not actually measure affect misattribution. The current study tested the validity of the AMP as a measure of affect misattribution by examining three components of the ERP known to be associated with the allocation of motivated attention. Results revealed that ERP amplitudes varied in response to affectively ambiguous targets as a function of the valence of preceding primes. Furthermore, differences in ERP responses to the targets were largely similar to differences in ERPs elicited by the primes. The existence of valence differentiation in both the prime‐locked and the target‐locked ERPs, along with the similarity in this differentiation, provides evidence that the affective content of the primes is psychologically registered, and that this content influences the processing of the subsequent, evaluatively ambiguous targets, both of which are required if the priming effects found in the AMP are the result of affect misattribution. However, the behavioral priming effect was uncorrelated with ERP amplitudes, leaving some question as to the locus of this effect in the information‐processing system. Findings are discussed in light of the strengths and weaknesses of using ERPs to understand the priming effects in the AMP.