Traditional and Network Measures of Peer Group Structures Reveal Differences Between Cultural and Socioeconomic Contexts

• In Event: 2-070 - Considering the Impact of Social Context and Group Norms Across Cultures

Integrative Statement

Introduction: Network analysis techniques and traditional measures provide a means of examining and understanding the differences in peer group structures of children from lower- and upper-middle-class families and from different cultural contexts. In this study, key structural indicators of children’s same-gender peer groups were measured at the level of the individual and the group. Individual level indicators individual included sociometric measures of acceptance, peer assessments of care, prosocial behavior, and popularity, and two network-based measures of centrality, specifically betweenness, which indexes a person’s central place among peers, and eigenvector centrality, which measures the degree to which a child is linked to peers who have many links to other network members. Group level indicators were group means of acceptance, care and prosocial behavior, and the network-based measure of cohesion. Two cultural contexts were studied: Montreal, Canada and Barranquilla, a city in the Caribbean region of Colombia that is known to have a collectivist culture.

Participants and Methods: Fifth- and sixth-grade boy and girls (N = 357) from 18 different classrooms in four schools Montreal and Barranquilla completed an unlimited choice nomination-based sociometric measure and a peer-assessment measure at a single time in the fourth month of the school year. The classrooms groups were cleaved by gender to create 36 same-gender peer groups which were used as the child’s peer network. Acceptance, centrality and cohesion scores were derived from the sociometric data; measures of care, popularity, and help were derived from the peer assessment data.

Results: Analyses revealed the following:

1. Consistent with our hypotheses, peer groups from Barranquilla were observed to show higher levels of within-group liking and cohesion.

2. Group means for prosocial behavior and care were positively correlated with the group-level measure of cohesion. Group-level cohesion accounted for the higher levels of prosocical behavior seen in groups from Barranquilla and from upper-middle-class groups.

3. Place and SES differences were observed with the centrality measures. Betweeness was higher in Barranquilla than Montreal and for lower middle class children; eigenvector centrality was higher in Montreal than Barranquilla and for upper middle class children. More importantly, in Barranquilla, care, help, and popularity were more strongly associated with eigenvector centrality (rs = .41, .33, and .25) than to betweenness (rs = .03, .02, and -.08) whereas in Montreal they were more strongly associated with betweenness (rs = .28, .33, and .19) than eigenvector centrality (rs = .05, .08, and .03).

4. Measures of centrality were associated with care, help, and popularity more for the lower middleclass children (rs = .28, .25, and .21 for betweeness; rs = .39, ,33, and .21 for eigenvector centrality) than for upper middleclass children (rs = .14, .19, and .09 for betweeness; rs = .19, .21, and .24 for eigenvector centrality).

Conclusion: Findings indicate that peer experiences vary substantially across cultural contexts and SES. The significance of measures of centrality differs as a function of SES. Subsequent analyses will assess contextual differences in the effects of structural differences on well-being.

Authors

• William M. Bukowski, Concordia University
  Presenting Author

• Jonathan Santo, University of Nebraska Omaha
  Non-Presenting Author

• Ryan Joshua Persram, Concordia University
  Non-Presenting Author

• Melisa Castellanos, Concordia University
  Non-Presenting Author

• Luz Stella Lopez, Universidad Autonoma del, Caribe
  Non-Presenting Author