Using Instrumental Variables to Account for Selection Effects in Research on First-Year Programs

The widespread popularity of programs for first-year students is due, in large part, to studies showing that participation in first-year programs is significantly related to students’ academic success. Because students choose to participate in first-year programs, self-selection effects prevent researchers from making causal claims about the outcomes of those programs. The present research examined the effects on first-semester grades of students participating in themed learning communities at a research university in the Midwest. Results indicated that membership in themed learning communities was positively associated with higher grade point averages, even after controlling for entering ability, application date, gender, and first-generation/low-income status. However, when instrumental variables were introduced to account for self-selection, the effects of themed learning communities on grades were not statistically significant. The results have implications for campus leaders and assessment practitioners who are working to develop methods for understanding the effects of programs designed to enhance the undergraduate educational experiences on their campuses.