Section III Applications
Personal Attitudes, Perceived Attitudes, and Social Structures: A Social Selection Model
Dean Lusher and Garry Robins
Perceptions of Others and Social Behavior
An important insight of social network analysis is that social action may arise due to actor characteristics (Emirbayer & Goodwin, 1994; Kilduff & Krackhardt, 2008). The process of social selection specifies an interaction between social relations and actor-level attributes. Examples include gender homophily (McPherson et al., 2001), network closure and psychological predispositions (Kalish & Robins, 2006), and delinquency behaviors (Snijders & Baerveldt, 2003). Furthermore, interdependency between social relationships and identity was argued by White (1992) and implied by social identity theorists (Tajfel & Turner, 1979; Turner et al., 1987).
Apart from the personal characteristics of network actors, social action may also be motivated by the cues people take from others around them in the social setting. Previous work on perceptions and social networks focused on the perception of social ties (e.g., “cognitive social structures”; Krackhardt, 1987). Also, both social position in the network and individual personality differences were shown to be important for the accurate perception of the network (Casciaro, 1998). In this chapter, however, we discuss the effects of individual perceptions of the attributes of others in the network rather than the perception of ties. The attributes under consideration are perceived attitudes. To delineate terminology clearly, we note a fundamental difference between personally held attitudes (“personal attitudes”) and an individual’s perception of the attitudes of others (“perceived attitudes”). In the study described in this chapter, perceived attitudes are based on individual perceptions of the attitudes “generally held” in the group.
The importance of perceived attitudes on the behavior of the individual is well supported by a number of social psychological theories. For instance, “groupthink” (Janis, 1971) refers to conforming to or conferring with group beliefs. This group consensus may override realistic situation appraisals and/or the individuality of group members, resulting in suboptimal decisions and outcomes for the group and its individual members. “Social comparison theory” (Festinger, 1954) suggests that in ambiguous situations or where objective criteria are lacking, people assess the veracity of their own views through comparison with the attitudes of others around them. Social comparison theory highlights how perceptions of the attitudes of others may guide personal attitudes, and therefore behavior, in part because they can resolve uncertainty. “Pluralistic ignorance” (Katz & Allport, 1931) - where most of a group’s members do not endorse a particular norm but incorrectly assume that it is accepted by most others - is also relevant. In essence, pluralistic ignorance is about “systematic errors in norm estimation” (Prentice & Miller, 1993, 244), although the “illusion of support is validated when it motivates widespread public compliance” (Centola, Willer, & Macy, 2005, 1010). Perceptions of the attitudes of others have regulatory effects on people’s behaviors, in part because people fear social sanctions (Centola et al., 2005).
The preceding social psychological theories indicate the importance of considering both personal and perceived attitudes with regard to social action. It is plausible, then, that personal and perceived attitudes may jointly affect social tie formation, and a notable limitation of the social psychological theories discussed previously is that they tend not to consider social relationships, either theoretically or empirically.
A theoretical perspective through which social relationships, personal attitudes, and perceived attitudes may be integrated is provided by “social identity theory” (Tajfel & Turner, 1979) and the theoretically related “self-categorization theory” (Turner et al., 1987), jointly referred to as SIT. SIT is a social psychological theory of group processes, intergroup relations, and norms. Social groups are cognitively represented in terms of prototypes, a subjective representation of the beliefs, attitudes, and behaviors definitive of a social category (Hogg, Terry, & White, 1995). Prototypes emphasize the similarities within and the differences between groups. Group identification occurs because “self-enhancement is best achieved through strategies that achieve a sense of in-group superiority relative to an out-group” (Hornsey & Hogg, 2002, 203). Those within the group who best exemplify the prototype are the most admired. It is important to note that the prototype is not the “average” in the group but is more extreme, a standard or norm to which group members may aspire. SIT suggests that the norm may reflect the attitudes of the powerful, the “more extreme individuals” and “their individual attitudes or characteristics.” SIT therefore proposes that conceptualizing group attitudes as the average of all individual attitudes is not sensible because power is not evenly spread throughout a group. Instead, SIT indicates that social norms more likely reflect the personal attitudes of powerful people in the group who have “a common interest in creating categories and stereotypes which are favorable to their power position and their social identity” (Lindenberg, 1997, 304).
From a social network perspective, the idea that some individuals in a social network may be more prominent than others is well accepted (e.g., “centrality” implies the importance of prominent actors within a network - see Freeman, 1979). Our research goal is to combine the theoretical insights of centrality and SIT to examine how perceptions of the attitudes of others relate to social network ties. Our argument is that group norms are related to the individual attitudes of popular individuals in the group. In line with SIT, we propose that individuals exemplify the group identity by holding certain extreme prototypical attitudes. Other actors, who themselves may not hold such extreme personal attitudes, perceive the prototypical attitudes as generally held among the group and acknowledge the prominence of those individuals who exemplify them. As an outcome, social ties tend to be directed toward these leading individuals, who thereby become more central in the network. To examine this proposition empirically, we analyze actor-relation effects for both personal and perceived attitudes. We do this while controlling for purely structural self-organizing effects in our networks (e.g., reciprocity, transitivity).