Evaluating a Scalable Program for Undergraduate CS Research

Undergraduate research experiences have been shown to have many positive effects on undergraduates including increased confidence, sense of belonging and retention. However, many previous studies of undergraduate research experiences have focused on advanced undergraduate (juniors and seniors) in one-on-one research experiences with a faculty mentor. Less is known about the effects of early undergraduate research, particularly via opportunities that scale beyond one-on-one faculty-student relationships to encompass large numbers of early undergraduates. The research question addressed in this work is whether a more scalable group-based research model aimed at early undergraduates from groups underrepresented in computing would show the same kinds of benefits for participants as more personalized one-on-one programs aimed at more advanced students. We evaluated a group-based early research program in the computer science department of a large public university. Through survey data and direct measurements of performance and retention several years after students had completed the program, we found that students who participated in this program have higher overall GPAs, more confidence, and more interest in research compared to several different control groups. Our design also allowed us to examine the considerable impact that selection bias can have on the evaluation of research programs. This work both validates the scalable structure of this research program and provides a richer perspective on the benefits of early undergraduate research in CS.