Search

We investigate whether student participation in undergraduate research--a high impact practice at TBR community colleges--has an effect on (1) academic performance; (2) probability of graduation, university transfer, and student departure; and (3) time to completion, transfer, and departure. Using enrollment, graduation, and course-taking data on community college students from the Tennessee Board of Regents and National Student Clearinghouse, we track the 2017 freshmen cohort over twelve calendar semesters and examine their exposure to undergraduate research experiences and their key educational outcomes. We generate propensity scores with a machine learning algorithm and use a doubly robust inverse probability weighting estimator to mitigate the selection bias and compare outcomes of similar students among undergraduate research participants and nonparticipants. Overall, we find that undergraduate research participants demonstrate better academic performance, are more likely to graduate and transfer, are less likely to depart, and progress to transfer and departure slower than similar nonparticipants. The effects grow in size with an increase in frequency of undergraduate research experiences.

Motivation for the study: The Tennessee Board of Regents (TBR) is implementing and expanding a suite of high impact practices (HIP) in its community colleges. Service learning, one of such HIPs, is incorporated into general education or degree programs' requirements via credit-bearing service learning components of different durations. While service learning participation and student results are examined periodically, this study contributes to the program evaluation by conducting a series of quantitative analyses that aim to assess whether a causal relationship exists between service learning experiences and key educational outcomes of first-time freshmen at community colleges. Objectives: To investigate whether participation in serving learning HIP affects: (1) the probability of earning a college credential, transferring to university, and student departure; (2) time to graduation, university transfer, and departure; and (3) academic performance. To examine if and how these effects, if any, differ by service learning duration and frequency of participation. Methods: Different types of propensity score analysis were used, including inverse probability of treatment weighting and matching and nonparametric regression, to estimate the effect of service learning participation on outcomes of interest in general and by duration level. In the final models, propensity scores were estimated using generalized boosted modeling. Dosage analysis was employed to examine the effect of frequency of service learning participation. The following modeling techniques were used for different outcomes: logistic and OLS regression, and event history analysis (Cox regression model). Results: Service learning participation increases the probability of graduation and university transfer, decreases the probability of student departure, expedites progression to graduation, delays progression to university transfer and departure, and is associated with a higher final GPA. The estimated results differ by duration level and frequency of service learning experience. Conclusion: Service learning, as implemented at TBR colleges, is an efficacious HIP, which contributes to both community college freshmen student success and institutional performance.