Evidence of resistance in response to a rape attempt is critical for prosecution, and is assumed by jurors to be information that would convince them of an assailant's guilt. Furthermore, there is growing empirical evidence that resistance during an actual attack increases the likelihood that victims will avoid being raped. Yet, resistance information can be conceptualized as essentially gender stereotype-inconsistent information. One consequence of such stereotype-inconsistent evidence may be rather paradoxical-it increases attributional processing concerning the victim's actions, and as a result, may decrease the blame assigned to the assailant for the attack. In two experiments, one using an instance of stranger rape and another using an example of acquaintance rape, we found support for the prediction that too much resistance on the part of a victim increases sympathy for the rapist-decreasing observers' confidence that the assault actually was rape and decreasing the sentence advocated for him, while creating uncertainty concerning the victim's causal role in producing the outcome. While some evidence of resistance is necessary to convince judges that the event was rape, increasing amounts of resistance can ultimately exert the opposite effect. Discussion centers on how gender-based expectancies influence judgments of rape victims, and the mental processes responsible for such counter-intuitive effects.