Many active‐learning approaches have been shown to be effective across contexts; however, discrepant results raise challenging questions about specific implementations and their efficacies. Here, we present two studies that examine how different instructors implemented the same active‐learning approaches in large introductory biology courses (about 300 students) and what factors correlated with positive student outcomes. In the first study, we analyzed student outcomes in two sections of a course. The course was collaborative developed by two instructors (A and B), and the sections (A and B) shared all teaching material and had common exams. However, students in section A had more positive and equitable outcomes: Section A had higher learning gains measured by pre/post concept inventories (effect size = 0.89 vs. 0.72, p<0.05) and a higher completion rate (94% vs. 83%, p<0.01); completion rates are different for majority vs. underrepresented minority (URM) students only in section B (85% vs. 67%, p<0.05). In parallel, we dissected how the two instructors taught and triangulate their actions in the classrooms with their beliefs about teaching. Using COPUS, we found that in lectures with clicker questions, instructor A had significantly more two‐minute time intervals in which he answered questions (47% vs. 22%, p<0.05) or had extended discussions with students (20% vs. 0%, p<0.01). In lectures dedicated to small‐group problem solving, instructor A had more two‐minute time intervals in which he moved among groups (79% vs. 32%, p<0.01) and fewer two‐minute time intervals in which he waited at the podium (17% vs. 56%, p<0.01). Using an established interview protocol and coding scheme, we found that instructor A had an acquisition belief of teaching (i.e. helping students learn content), whereas instructor B had a transmission belief of teaching (i.e. disseminating information). These beliefs are consistent with the actions observed by COPUS and may provide an explanation for pedagogical decisions made either explicitly or implicitly by the two instructors. In the second study, we expanded our analysis to include a third instructor (C), who had a development belief of teaching (i.e. facilitating the discovery of knowledge by students in a constructivist manner), teaching a different course. We found that while instructor C has a mixed COPUS profile with elements similar to both instructors A and B, his group activities had a fundamentally different structure that guides students to make their own discoveries and formulate mental models. Together in these two studies, we examined how multiple people enacted the same curriculum, why they implemented active learning in specific but distinct manners, and what beliefs about teaching and actions in the classrooms correlated with positive student outcomes. From these results, we propose a new model to understand how and why faculty teach differently and what features of their teaching promote student success.This abstract is from the Experimental Biology 2018 Meeting. There is no full text article associated with this abstract published in The FASEB Journal.