Three studies examined the hypothesis that when perceivers learn of the existence of multiple, plausibly rival motives for an actor’s behavior, they are less likely to fall prey to the correspondence bias than when they learn of the existence of situational factors that may have constrained the actor’s behavior. In the first 2 studies, Ss who learned that an actor was instructed to behave as he did drew inferences that corresponded to his behavior. In contrast, Ss who were led to suspect that an actor’s behavior may have been motivated by a desire to ingratiate (Study 1), or by a desire to avoid an unwanted job (Study 2), resisted the correspondence bias. The 3rd study demonstrated that these differences were not due to a general unwillingness on the part of suspicious perceivers to make dispositional inferences. The implications that these results have for understanding attribution theory are discussed.