To strengthen the US scientific workforce, we aim to recruit and retain talented students in science, technology, engineering, and math (STEM) fields, to enhance success among students from groups underrepresented in STEM fields, and to diversify the scientific workforce to mirror the US population. Given that opportunities for authentic research may support STEM advancement, we seek to maximize the number of students involved in research. Our Behavioral Research Advancements in Neuroscience (BRAIN) research program tests the hypothesis that a team-based collaborative-learning model not only provides research opportunities for more students, but also produces outcomes at least as positive as a traditional one-on-one apprenticeship model. We examined scientific research self-efficacy as a critical construct for measuring student outcomes and predicting student progress toward STEM careers. Here we provide descriptive portraits of four women who participated in BRAIN, integrating quantitative survey data with analysis of pre-and post-semistructured interviews. Although selected for different self-efficacy trajectories in the quantitative surveys, all four women described increased self-efficacy in interviews and emphasized mastery experiences as a source of self-efficacy. Two women illustrate one general outcome from the program: women overcame initially lower scientific research self-efficacy, matching self-efficacy among men by mid-program. The overarching study suggests that both team-based research and apprenticeships can raise scientific research self-efficacy, which predicts STEM career success. Therefore, this collaborative model provides a structure for authentic research at institutions that may lack available mentors, and yet aim to improve opportunities for diverse undergraduate groups to pursue STEM careers.