Board 158: Engagement Patterns Across Race, Gender and Family Income in Engineering Classrooms

Abstract

This IUSE-funded study investigated differences in behavioral and emotional engagement that emerge across family income, gender, and race in engineering classrooms. Engagement levels and engagement patterns were measured across seven sophomore-and junior-level engineering courses at a large public university. Differences in engagement were evaluated quantitatively between the two numerical majority races in this study (Asian, white), between genders, between international and domestic students, and across three levels of family income. Sample sizes for other racial groups (black, Pacific Islander, Native American, Hispanic, and Other) were too small to support analyses by family income and were not included in this study. Initial analyses of variance (ANOVA) revealed significant differences in at least one form of engagement between Asian and white students, between men and women, between domestic and international students, and across family income levels. As a result, all four demographic variables (race, gender, country of origin, family income) were retained in a subsequent linear regression to understand potential interactions among these demographic variables. Since these models were weak, the analysis then looked at engagement patterns rather than engagement levels. In this next phase of analysis, scores for the five engagement variables were classified using a non-parametric k-means clustering approach. The data optimally separated into two main categories: less engaged students (Cluster 1) and more engaged students (Cluster 2). Among domestic students, 100% of low income Asian women and 82% of low income Asian men (82%) fell into the more engaged cluster, while high-income Asian women (83%) fell into the less engaged cluster. Among international students (who were entirely Asian in this sample), low income Asian men and high income Asian women were among those who had the highest percentage of lesser engaged students (40% of each group, respectively) while middle income Asian men and middle income Asian women had the highest percentage of more engaged students (approximately 80% of each). Overall, the k-means clustering approach provided greater insight into the data than traditional statistical analysis techniques. Differences and trends among all four demographic variables (gender, family income, race, country of origin) emerged, showing that students from some demographic groups seem more susceptible to remaining less engaged in courses than other groups.