Working Paper: NBER ID: w10370
Authors: Eric Bettinger; Bridget Terry Long
Abstract: One of the most pronounced trends in higher education over the last decade has been the increased reliance on instructors outside of the traditional full-time, Ph.D.-trained model. Nearly 43 percent of all teaching faculty were part-time in 1998, and at selective colleges, graduate assistant instructors teach over 35 percent of introductory courses. Critics argue that these alternative instructors, with less education and engagement within a university, are causing the quality of education to deteriorate and may affect student interest in a subject. However, little research exists to document these claims. This paper attempts to fill this void using a unique dataset of students at public, four-year colleges in Ohio. The paper quantifies how adjunct and graduate assistant instructors affect the likelihood of enrollment and success in subsequent courses. Because students with alternative instructors may differ systematically from other students, the paper uses two empirical strategies: course fixed effects and a value-added instructor model. The results suggest that adjunct and graduate assistant instructors generally reduce subsequent interest in a subject relative to full-time faculty members, but the effects are small and differ by discipline. Adjuncts and graduate assistants negatively affect students in the humanities while positively affecting students in some of the technical and professional fields.
Keywords: Adjunct Instructors; Graduate Assistants; Student Success; Higher Education
JEL Codes: I2; H4
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| adjunct and graduate assistant instructors (A23) | subsequent interest in a subject (Y20) |
| adjunct and graduate assistant instructors (A23) | student outcomes (A21) |
| younger adjuncts (under 40) (A22) | negative effects on student outcomes (I24) |
| course fixed effects (C23) | control for unobserved heterogeneity in students' characteristics (C21) |
| value-added model (D46) | assess impact of instructor quality on student outcomes (I24) |
| Heckman selection models (C24) | control for selection biases (C90) |