Comparing Student Outcomes in Online vs. In-person Sections of an On-campus Computer Science Course

We present results from a quasi-experimental study comparing performance and sense of belonging between online (N = 212) and in-person (N = 116) sections of the same computer science class, where all students were residential and physically present on campus. When we view the class as a whole, we find that the performance of students on all formative assessments (pre-lecture activities, in-class activities, homework, coding projects) is statistically equivalent. While performance on exams is statistically significantly higher in the in-person section even after controlling for students' incoming GPA (2.4 percentage points higher when compared with students in the online section, p = 0.028), we find no statistically significant difference in students' final weighted average score (the advantage for in-person students reduced to 1.2 percentage points with p = 0.067). When we disaggregate by gender, we find that women in the online section outperform the women in the in-person section. Men's, underrepresented minorities', and first-generation students' performance follow the same trends as the whole class. Sense of belonging was measured by surveys at the beginning and end of the course. Students were asked to respond to questions about their perceived comfort in the classroom, perceived isolation, and perceived support from course staff and other students. We note that the whole class's sense of belonging statistically increases from the beginning to the end of the semester in both sections. Furthermore, the increased sense of belonging is more pronounced in the in-person section. Based on our findings, we conclude that online sections for on-campus students may be an effective way to accommodate large class sizes, increased enrollment pressure, and students' need for flexibility, while not disadvantaging students' learning outcomes.