The analysis of personal attitudes and perceived attitudes with social network relations demonstrates some interesting results in these two local contexts. In the school, students who do not personally hold high domina- tive masculine attitudes send many social ties. In addition, students who perceive that others do hold high dominative masculine attitudes also send many ties. Furthermore, students who hold high personal masculine attitudes receive the most social ties (i.e., are popular individuals). We suggest that some form of social hierarchy may explain these results. SIT assumes that the norm (or prototype) is more extreme than the average attitude (Hogg et al., 1995) and representative of the qualities of the leaders of the group (Lindenberg, 1997). In combination, these effects suggest that those who perceive high dominative masculine attitudes send many ties to others in the network potentially because the personal attitudes of those who are popular (and thus the leaders) are also high. Social norms (measured here as perceived attitudes) are dictated by the attitudes of prestigious people of a social setting, affecting the formation of social ties.
However, in the football team, we have a different story in that perceived attitudes have an independent explanatory capacity regarding social tie formation. We note that masculine attitudes include a component about violence, and that we might anticipate those most demonstrating aggression to be high in personal MAI attitudes. Yet, aggression toward others within this team is not associated with personal attitudes to masculinity; rather, perceived attitudes are significantly associated with aggressive behaviors. Specifically, those who were aggressive to others were high in perceived MAI, as were those who were the victims of aggression. Thus, it is principally the perception by individuals of the attitudes of those around them that may push players to engage in aggression relations. The notion that “culture is a ‘tool kit’ for constructing ‘strategies of action’” (Swidler, 1998, 176) may well explain why players without a strong personal endorsement of violence, but with perceived attitudes that aggression is valued, are more likely to be aggressive toward others. The results for the team, based on a negative affect social tie, are therefore different than the results for the boys based on a positive affect network. However, the results of the footballers suggest that the “misperception” of social norms implied by the notion of pluralistic ignorance is also a feasible explanation.
These effects for perceived attitudes raise a fundamental question of whether social behavior, including tie formation, is related to personal attitudes, perceived attitudes, or some interaction of the two. Frosh et al. (2001) stated that boys in schools present different behaviors, depending on whether they are engaging with boys, girls, or adults. The presentation of contradictory masculine identities in these different scenarios is hardly likely to result from changing personal attitudes of the boys from one context to the next. Instead, the embodiment of different masculine behaviors is likely to be related to an individual’s differing perceptions of the attitudes of others in the context - that is, something we conceive of as an individual’s idea of the social norm. These perceived attitudes certainly interact with the personal attitudes of the individual and the specific social relations he has with others. We address these issues in terms of a social selection model, which makes theoretical sense from SIT. However, it is also possible that the personal attitudes of prominent (i.e., central) individuals may influence the perceived attitudes of peripheral network members. An alternative analysis using a social influence-type model such as autologistic actor attribute model (ALAAM: Chapter 9) could be used, but of course the direction of effect cannot be fully resolved without the use of longitudinal data.
Finally, the findings for the footballers suggest that certain contexts may invoke behaviors and relations (i.e., aggression) through expectations or norms, and personal attitudes may not be directly related to social behavior. Further theorizing about specific ways in which perceived attitudes may operate in relation to personal attitudes and local social structures may provide an enriched understanding of the processes that people engage in within their social worlds.