Developing an instrument of classroom social engagement

Abstract

Student engagement plays a pivotal role in the formation of engineers because it is central to many significant predictors of the academic and career success of engineering students. Research has shown that the kind of social interactions that students maintain with peers and instructors within their academic community influence their connectedness and social capital. Students' social capital impacts their access to resources that are vital to academic success, and may eventually affect their sense of belonging, resilience and grit, and the need to put in the effort needed for academic career success. Besides the importance of social networks and capital to students' engagement and academic achievement, research on social engagement has received renewed interest lately due to advances in methodologies of social network analysis. These studies are enabled by measurement instruments that purport to assess indicators of social engagement in the classroom. Although increased interest in social engagement is commendable, the data driving this research has been mostly reliant on analysis of the frequency of social interactions observed among students. While such analyses provide useful metrics for understanding social engagement, they are limited in helping researchers understand the underlying reasons for these engagement activities. To fill this gap, we are developing an instrument of social engagement that uses constructs of social capital, engagement and social networks to explicate the richer context of student social engagement as it relates to their interpersonal interactions with peers and faculty. In the current proposal, we present a psychometric study of a section of the instrument. Five hundred and thirty-four participants completed a social network instrument that assessed their interaction with peers, graduate assistants, and instructors based on 11 items created to assess three constructs of social engagement: value, reciprocity and conduit of belonging. Exploratory and confirmatory factor analyses were conducted to examine the structural validity and reliability of the hypothesized construct. An initial factor analysis showed that our data did not support the three factor-model we hypothesized (CFI =.841, RMSEA =.161). The final model indicates that the data from students' responses only supported a 2-factor model with 7 items (CFI =.986, RMSEA =.089). Internal reliability of the two scales were.94 and.89 Cronbach's alpha. The full paper will discuss scale revision.