Allport's (1954) contact hypothesis had two parts. Allport proposed that social contact with members of negatively stigmatized social groups would, under favorable circumstances, lead to both a) positive attitudes toward the specific group member(s) with whom the contact occurred and b) less negative attitudes toward the social group as a whole. The first part of Allport's contact hypothesis has been thoroughly investigated. Literally hundreds of studies have established that contact produces positive attitudes toward the specific group member only when the contact situation involves equal status, mutual goals, and active cooperation. The second part of Allport's contact hypothesis has received little empirical attention or support. With the exception of two studies, the few attempts to find generalization o specific to more general attitudes have suggested that generalization is unlikely even under ideal conditions. Positive contact that involves equal status, mutual goals, and active cooperation often leads to liking for the specific group member(s), but with no generalization to other individual group members or to the group as a whole. A comprehensive review of the contact hypothesis literature concludes that """"""""the failure to generalize positive attitudes to outgroup members in general is a critical weakness in traditional contact theory"""""""". It is important to investigate the precise conditions under which positive contact with one former mental patient, one former substance abuser, or one person with AIDS (to use three health-related examples) ameliorates or does not ameliorate general attitudes. Given the well-known propensity for self-fulfilling prophecies in social interactions, it is important to discover the types of positive contact experience that will alter stereotypes and """"""""carry over"""""""" into less prejudiced treatment of other members of an initially stigmatized social group and to define the ways in which such attitude- ameliorating experiences might differ from equally positive contact situations in which the specific person is well-liked, but without any change in attitude or behavior toward the larger social group. The proposed studies constitute the first systematic program of research to concentrate exclusively on the second part of Allport's contact hypothesis -- the generalization part. The proposed studies examine three factors that may affect generalization within the contact hypothesis: perceived typicality of physical appearance, salience of the specific member's connection with the group, and how much the information exchanged during contact separates the specific member from the group. All three factors are thought to operate because they heighten versus contradict the extent to which the specific member is perceived as a """"""""good"""""""" versus a """"""""poor"""""""" group representative.