On institutional belongingness and academic performance: mediating effects of social self-efficacy and metacognitive strategies

ABSTRACT

Educational researchers have provided evidence that students who see themselves as valued members of their university (institutional belongingness) tend to have higher academic performance than students with a weaker sense of institutional belongingness. The current research draws on social cognitive theory to inspect two mechanisms that might explain this correlation: social self-efficacy and metacognitive strategies. We tested a double-mediation model with a large sample of students (n = 1,480) from one higher education institution in New Zealand. Using structural equation modeling, social self-efficacy and metacognitive strategies were meaningful contributors to the relation between institutional belongingness and Grade Point Average (GPA). Our discussion focuses on how universities can design strategies that promote belongingness and, in turn, improve how students interact, learn, and perform.

KEYWORDS:

• Belonging
• performance
• social self-efficacy
• metacognition
• university

Acknowledgments

My thanks to the student participants and institutional research team. Additional thanks to the Studies in Higher Education editorial team and the anonymous reviewers for their constructive feedback in previous versions of this manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Brubacher, McMahon, and Keys (2018) reported a moderate, positive association between social self-efficacy and school belonging; however, their study represented a sample of 98 African American and Latinx adolescent students with an intellectual or physical disability. It would be reasonable to expect individual and group differences between their results and those from a general student population in higher education.

2 Absolute indexes were chi-square (χ2), a test of the difference between observed and estimated covariance matrices (p-values should ideally be greater than .05), and Root Mean Square Error of Approximation (RMSEA) with its confidence intervals. RMSEA values and confidence intervals should be less than .10 for acceptable fit and values less than .06 for good fit (Schreiber et al. 2006; Schumacker & Lomax, 2012). Relative indexes were Comparative Fit Index (CFI), a population measure of model misspecification, and Tucker–Lewis Index (TLI), representing relative fit by considering model parsimony. For relative indexes, values greater than .90 indicate acceptable model fit, whereas values greater than .95 indicate good model fit (Schreiber et al. 2006).

3 The direct effect in which the four covariates was also weak, positive, and statistically significant (B = .427, SE = .097, β = .131, p < .001).

4 Importantly, however, it is also worth emphasizing that this coefficient of determination explains a reasonable proportion of variance in Year GPA, but this estimate is attributed to both the predictor (X) and the mediators (M1 and M2), but neither alone (Fairchild et al. 2009; Lachowicz. 2018; MacKinnon 2008).