Stereotypes sometimes arise from faulty human memory.
If there were a social cognition hall of fame, Hamilton and Gifford's (1976) article on illusory correlations and stereotypes would surly be a featured exhibit. Offering an insightful analysis and a telling pair of experiments, they argued that stereotypes sometimes arise not from base human motivations or intergroup conflict but rather from the faulty workings of human memory. Hamilton and Gifford argue that negative behaviors are numerically rare. Thus, when performed by minority group members, who be definition are also numerically rare, these behaviors become disproportionately memorable, leading to later impressions that minority group members are responsible for more than their fair share of undesirable behavior.
Of critical importance to Hamilton and Gifford's argument was that no prior expectation regarding the rare group was necessary to produce their effects, just the unique memorability of rare-group/rare-behavior pairings due to their relative infrequency. … [P]articipants thought that members of the minority group were more likely to exhibit the rare behavior than members of the majority group.
… Subsequent research supported the contention that these sorts of illusory correlations are due to the enhanced availability of jointly distinctive information.
… The core of Hamilton and Gifford's analysis is that illusory correlations arise because rare-rare groupings are disproportionately available and more likely than other pairings to be recalled. The illusion arises, in other words, because of the impression that one has witnessed more examples of negative behavior committed by minority group members than was actually the case.
We suggest, however, that there are other forms of distinctiveness that can give rise to illusory correlations that associate rare groups with rare behaviors. We argue that one does not need to survey a series of behaviors performed by common and rare groups for an illusory correlation to emerge. Instead, just one unusual behavior performed by one person from an unfamiliar group is sufficient-leading to a phenomenon we term one shot illusory correlations.
One-shot illusory correlations arise because people often bring knowledge about the prevalence of various human behaviors to bear when witnessing an action performed by another person. Thus, based on a lifetime of experience, a person can code just a single instance of behavior as common or rare. People are likely to conclude that reading a newspaper or eating a sandwich are common events, whereas wearing a newspaper or kicking a sandwich are unusual. Rare behavior is generally more arresting than common behavior-and more likely to prompt thought and elaboration. Furthermore, a lifetime of interaction yields knowledge about the frequency with which members of different groups are encountered. To someone in the United States, an American would be coded as common but a Sri Lankan would be coded as rare. To Sri Lankans, however, the individuals considered common and rare would be reversed.
We contend that distinctive behavior performed by a member of an unusual group is not only arresting but also likely to trigger thoughts that lead to an association between the pertinent group and the behavior. Such associations arise because rare-rare pairings prompt people to entertain the hypothesis that the person's group membership might explain his or her unusual behavior.
… Illusory correlations may emerge not simply because distinctive behaviors performed by minority group members are better remembered byt because they instigate an attributional process in which group membership is considered a possible explanation of the unusual behavior.
Source: One-Shot Illusory Correlations and Stereotype Formation