Social Capital as Predictor and Mediator of the School Engagement, Trouble Avoidance and Academic Performance of Middle School Students

Background and Purpose: A growing body of research has found socially supportive relationships in the lives of youth promote positive school outcomes. The conceptualization of the influence of such supportive relationships fits within a social capital perspective and has been found to influence school behaviors, school engagement, and academic performance. However, much research is needed to understand the complex and cumulative influence of such social capital across the central microsystems that impact youth. The purpose of the current study was to investigate whether groups of students could be identified according to patterns in their experiences of social capital across those microsystems—home, school, peers, and neighborhood. Further, we examined whether those patterns would then explain variation in school outcomes. The current study was completed with data collected with the School Success Profile (SSP), an ecological survey that assesses social capital assets across those microsystems.

Methods: In the context of student demographics, we used a four-step analytic strategy to investigate associations between social capital and school outcomes. First we used latent class analysis (LCA), a mixture modeling approach, to define student groups based on the probabilities observed on indicators of social capital. LCA defined a latent discrete condition identifying unique combinations of social capital using seven SSP social capital dimensions: neighbor, family, friend and teacher support, family togetherness, peer group acceptance and social support. After labeling identified classes according to the item probabilities within each class, a stepwise regression of three school outcomes was conducted. First, demographics were entered; second, free/reduced lunch was added (proxy for income). Third, the classes identified in the LCA were added. Hierarchical linear modeling was used to account for the nesting of students within schools, and multiple imputation was used to address missing data.

Results: Five classes of social capital emerged from the LCA: high (overall), high family, high friend and teacher, high teacher, and low (overall). In the stepwise regression, free/reduced price lunch decreased significant negative associations for Latinos (on trouble avoidance) and African Americans (on grades) to non-significance. When social capital was added, these demographic differences were reduced further, and positive associations for African American and Latino students with school engagement were strengthened. Students with high overall social capital revealed the best outcomes relative to those having low overall social capital. However, high teacher and friend social capital were also strongly protective relative to low overall social capital.

Conclusions and Implications: Developing educational interventions based on increasing the social capital in the lives of students holds promise to advance the school success of all students. The current study informs such efforts by revealing the cumulative impact of social capital across settings. Further, the current findings emphasize the higher impact supportive social capital can have for poor students and students of color, who both showed reduced group differences in school outcomes and increased promotive effects when social capital latent class membership was added to our models. Furthermore, the most protective types of social capital—friends and teacher—can be directly influenced within the school setting.