Association for Public Policy Analysis & Management

Policymakers have urged for a focus on improving the rates at which students enter and complete four-year STEM degrees. One such product of these general efforts is disciplined-based education research (DBER) – research that aims to study and promote the adoption of disciplinary specific classroom instructional techniques that will increase the quality of teaching and learning in STEM education (coined “promising practices”) (National Research Council, 2012). In addition to these efforts, the hiring of faculty dedicated to teaching—that is, instructors with dramatically decreased research loads—has been on the rise at predominantly large research institutions. In this study, we leverage the dimension scores of an underutilized data-reduction technique to estimate how promising practices are moderated by instructor characteristics.

This study uses a data set of nearly 300 classroom observations for a wide range of STEM courses, including those in biology, chemistry, engineering, information and computer sciences, math, physics, and statistics. There were 86 instructors—some of whom were observed more than once, either for the same course or a different one—who’s classroom teaching techniques were observed and fully recorded. We restricted our sample to: (1) students who were observed in the dataset at least 3 or more times, (2) students who did not repeat the same course during the time of observation, and (3) students who had complete data on course final grade (listwise deletion). The total number of student observations in our data were 35, 937. And the unique number of students (head count) in our sample was 7,679. The primary independent variables in this study are the five teaching dimensions extracted from the multiple correspondence analysis. Two instructor characteristic variables were used to moderate the impact of the promising practices on students’ grade in the course, gender and instructor rank. Gender was dummy coded as a binary variable with females as the reference group. Instructor rank was defined as the primary job function of the instructor—that is, what an instructors’ job performance is primarily based upon. Three categories were generated: a research incentive category, teaching incentive for full-time instructors, and teaching incentive for part-time or adjunct instructors.

In this study, we use regression analysis to understand how the predictive ability of the promising practices on students’ grade in an introductory STEM course is moderated by instructor characteristics. Student and course fixed-effects are employed when specifying the interaction model. The results suggest that teaching faculty employing higher levels of two of the promising practices has a positive influence on student outcomes. The implications of this research suggest that the incentives instructors have in teaching relate both to the pedagogical decisions they make and the consequences it has on student learning. Policy recommendations are discussed with respect to the hiring of teaching-focused faculty. Future research should replicate and extend the work from this study by investigating next grade in the STEM sequence to estimate the effects of instructor characteristics and practices on more distal outcomes.