STEM Enrollment Decision Trees as Graduation Predictors for Community College Students Enrolled in Remedial Mathematics

Abstract

Objective/Research Question: Community college graduation rates are typically quite low, and developmental mathematics enrollment and coursetaking patterns may constrain academic outcomes. To identify ways in which community college graduation rates may be improved, decision trees were utilized to examine the STEM coursetaking patterns of N = 5,065 students who matriculated in remedial mathematics. Methods: The research design was guided by Tinto’s academic and social integration framework, which provided an analytical lens for identifying how decision trees facilitate academic decision making when academic and social integration is limited. Decision trees identified course sequence rules to predict graduation, which can be used to formulate course pathways for community college advisors and their students. Results: Nine rules from the decision tree were identified, which could be used to advise community college students in coursetaking that aligns with career aspirations. The most important variable predicting graduation was completing College-Level Mathematics, which included Algebra II, Statistics, Precalculus, and survey mathematics courses. General education sciences courses such as Astronomy, Geology, Environmental Science, and Marine Biology were the most important science courses predicting graduation. Conclusions/Contributions: Results suggest the importance of College-Level Mathematics in providing the skills necessary for students to be successful in subsequent STEM coursework and persist to graduation. Designating specific academic pathways may improve social and academic integration and graduation rates, providing continuity as students work with different advisors to choose majors and plan course sequences. Transparent, accessible enrollment planning fosters programmatic consistency and student agency in selecting coursework that will maximize their success.