We investigated social-cognitive outcomes for U.S. university students engaged in a supplemental small-group peer-learning program which provided a more intimate learning environment than the associated course. We surveyed 604 students taking part in such a program at the beginning and end of an academic term on social-cognitive factors related to their experience in an introductory-level STEM (science, technology, engineering, and mathematics) course, to which the supplemental peer-learning program was attached. A comparison group consisted of 676 non-participants enrolled in the same courses, but not taking part in the supplemental program. Those who had participated in the peer-learning program experienced greater gains, relative to the comparison group, in self-efficacy for course tasks and self-regulation for learning, and they were less likely to use surface-level memorisation in studying for the course. The findings suggest that small-group learning in the university setting can be used as a means to improve students' study habits and confidence about the course, contributing social-cognitive outcomes to an existing body of literature which emphasises academic performance outcomes.