2026 ASEE Annual Conference & Exposition

Securing an internship during undergraduate studies is a valuable milestone for computing students, as it provides students with hands-on experience, enhances their technical skills, and increases their employability for full-time job opportunities. However, we have limited knowledge about our ability to identify which students in computing programs are likely or unlikely to participate in internships. To address this limitation, we conducted a survey-based study to examine predictors of internship participation among 512 undergraduate computing students enrolled at three universities in the United States. We previously modeled the data, but in this study we evaluated the efficacy of four common predictive models - K-Nearest Neighbors, Random Forest, Logistic Regression, and Neural Networks - to understand student participation in internships. Our findings indicate that the Random Forest and Logistic Regression algorithm achieved the highest accuracy in predicting internship participation at 71%, while the K-Nearest Neighbors and Neural Networks model performed the worst, with an accuracy of 64%. Our results provide baselines for future models and contribute to a better understanding of factors predicting internship participation. Additionally, they can inform academic institutions and career services in supporting students’ professional development.