Practical and Theoretical Implication of Different Analyses on Friendship Nominations, Peer Acceptance, and Rejection

• In Event: 1-219 - Considerations in the Assessment of Social Context and Peer Network Structure

Integrative Statement

Rationale: The nuances of explaining differences in popularity, acceptance, and rejection in childhood and adolescence has been the topic of robust discussion (Rubin, Bukowski & Parker, 2008). More recently, advanced statistical approaches (i.e.: actor-partner multilevel modeling and social network analyses) have allowed researchers to better disentangle the subtle factors which differentiate why some children are friended, liked or disliked more than others. Nevertheless, the advantages and disadvantages of these various approaches remain opaque. The current paper examines different ways of measuring friendship nominations, peer acceptance and rejection using gender, age and peer-assessed popularity, relational and physical aggression as predictors.

Method: The sample consisted of 430 early adolescents (meanage = 10.90 years, SD = 0.72; 51.6% male) from three schools in Montreal, Canada (participation rate of 87.58%). Participants provided same-sex peer nominations of popularity relational and physical aggression. Classroom-based friendships were also measured in addition to individual liking ratings of all classroom peers. We explored associations between traditional assessments of friendship nominations, peer acceptance and rejection with gender, age, and peer assessed popularity, relational and physical aggression as predictors (using multiple regression and then structural equation modeling). We then employed social network analyses again using gender, age, and peer assessed popularity, relational and physical aggression as factors to explain differences in indegree ties. Finally, we used a three-level logistic multilevel framework nesting all child-peer dyadic ties within individuals nested within their classrooms.

Results: Analyses of the traditional measures of friendship nominations peer acceptance and rejection revealed associations in the expected directions (figure 1). There were few gender and age differences whereas early adolescents perceived as popular by their same-sex peers were more likely to be chosen as a friend, more accepted and less rejected. On the other hand, relationally or physically aggressive early adolescents were less likely to be chosen as a friend, less accepted and more rejected. These effects were subtly different in the social network analyses and MLM models but there were also additional effects which further explained differences in mutual ties.

Discussion: The results reveal more similarities than differences between the three approaches. Although traditional measures of friendships, peer acceptance and rejection are easy to calculate, they lack the ability to disentangle dyadic effects. Social network analysis on the other hand can masterfully model the effect of individual factors on being chosen by peers, reciprocated ties and other interesting details of the social network. It is limited, however, in requiring full network data and fails to account for classroom level differences. Finally, the multilevel modeling framework provides a very nuanced description of the effects of being chosen by a peer and reciprocation while also accounting for the characteristics of the classes. It cannot however account for all of the characteristics of a social network. We advocate for the adoption of MLM model that can simultaneously examine the between individual effects of the other approaches but moreover delineate how friendship nominations, peer acceptance and rejection are associated with each other concurrently and across time.

Authors

• Jonathan Santo, University of Nebraska Omaha
  Presenting Author

• Felicia Meyer, Concordia University
  Non-Presenting Author

• William M. Bukowski, Concordia University
  Non-Presenting Author