Association for Public Policy Analysis & Management

Due to decline in public financing for higher education, colleges have been motivated to increasingly rely on non-tenure track faculty partly by cost-containment and partly by flexibility since adjuncts, especially part-timers can be relatively easily increased or decreased. Yet, typically hired into temporary positions and most often with only part-time employment, adjunct instructors have been found to face various challenges in interacting with students regularly, providing important information, engaging and motivating students. However, due to data limitation, whether the limitations with temporary faculty may translate into any negative effect on student college persistence and longer-term labor market outcomes remains an unanswered question.

Our study begins to shed light on these issues by using a novel dataset that links student college transcript and enrollment records with detailed instructor information teaching each course in an entire state higher education system including all public four-year and two-year colleges. This administrative dataset is further linked to unemployment insurance data which enables us to track student labor market performance up to eight years after a student’s initial college enrollment. To examine the possible influence of different instructors on college persistence and labor market performance, we conduct a student-level analysis that relates the proportion of course credits taken with different types of instructors during a student’s initial year in college to his/her probability of withdrawing from college by the end of first year, as well as average annual earnings at the 6^th, 7^th, and 8^th year post initial college enrollment. To minimize bias from student self-selection into courses taught by different types of instructors, we use a course-set fixed effect model that compares between students who take exactly the same set of courses in their first year of enrollment; we further augment the course-set fixed effects model by combing it with an instrumental approach which exploits term-by-term fluctuations of faculty composition in each department as an instrument for the proportion of credits taken with different types of instructors, therefore controlling for both between- and within- course sorting.

Using detailed data on faculty contractual form, we are able to distinguish between temporary and long-term non-tenure track faculty and therefore divide all faculty into three groups: i) temporary adjuncts, ii) long-term non-tenure track, and iii) tenure-track/tenured. Our results indicate that both two-year and four-year colleges heavily rely on non-tenure track faculty: In two-year colleges, all first-year courses are taught by non-tenure track faculty, where around 55% of the total first-year course enrollments are with temporary adjuncts; in four-year colleges, around 43% of first-year course enrollments are with long-term non-tenure track faculty and 22% with temporary adjunct faculty. Results from the IV strategy with course-set fixed effects show an overall negative correlation between the proportion of first-year credits taken with non-tenure track faculty and students’ probability of persisting in college after their first year. Such negative association is particularly stronger among temporary adjuncts than long-term non-tenure faculty. However, no significant impact is identified in terms of earnings.