Researchers tested 281 undergraduates to determine if positive behavior messages about African American males presented during a learning task affected scores on explicit and implicit racial prejudice measures. During the learning task, we manipulated how many positive messages the participant viewed (100 vs. 150 or none) and the amount of African American males these messages applied to (1 vs. 3). Participants who viewed 150 positive messages about one African American male displayed more explicit prejudice than participants in control groups or participants learning 100 messages about one person. Results for the implicit measure indicated that participants who learned about three people and viewed 150 messages had faster implicit associations between African American males and positive adjectives when compared to participants who viewed fewer messages or learned about only one person. These findings demonstrate that learning positive information about a target group generalized to other exemplars from that category, but only when there was more than one example.