An oft-cited maxim in higher education is that ''faculty teach the way they were taught'' because they receive little formal training in teaching before entering the classroom. However, little is known about the origins of faculty knowledge about teaching or the role their prior experiences play in the development of their teaching practices. In this exploratory study, we interviewed and observed 53 science, technology, engineering, and mathematics faculty at three research institutions. Using qualitative analysis methods (i.e., thematic and causal network analysis), we find that faculty do not only model their teaching after previous instructors, but also draw upon a varied repertoire of knowledge and prior experiences. These include knowledge derived from their experiences as instructors (46 respondents), their experiences as students (22 respondents), their experiences as researchers (9 respondents), and from their non-academic roles (10 respondents). In-depth analyses of two faculty members elaborate on the relationship between these varied types of prior experiences and how they interact with other factors including beliefs about teaching, instructional goals, and features of the organizational context to ultimately shape their classroom practice. The results suggest that instead of assuming that faculty lack any knowledge about teaching and learning, professional developers and policymakers should instead acknowledge and build upon their preexisting ''craft'' knowledge as professional teachers. Future research should focus on relationships between specific types of knowledge and teaching practice and how these varied experiences influence identity formation.

Descriptions of faculty practice that illuminate nuances of how course planning and classroom instruction occur in specific contexts are important to inform pedagogical interventions. The study reported in this article draws on systems-of-practice theory to focus on the dynamic interplay among actors, artifacts, and tasks that constrains activities such as course planning and constitutes other activities, such as classroom instruction. This qualitative case study of faculty teaching in math and science disciplines at 3 research universities is based on interview and classroom observation data (n = 57 instructors) that are analyzed using causal network and social network analysis techniques. Results indicate that course syllabi are important organizational artifacts that are created by curriculum committees, inherited from previous instructors, and shaped by consideration of the sequential acquisition of knowledge. Faculty perceived that although course syllabi delimit the type and temporal sequencing of material for faculty, they are generally free to teach how they like. Observation data reveal discipline-specific configurations in frequently used teaching methods, cognitive engagements, and the use of instructional technology. These results also demonstrate that conceptualizing teaching solely as the use of particular methods (e.g., lecture) obscures subtle features of practice. Using the approach outlined in this article, instructional designers can obtain insights into meanings and practices that can be used to design and implement locally attuned reform initiatives.

Given the limited adoption of research-based teaching methods at the postsecondary level, research is necessary that examines why faculty choose to teach the way they do. In this article, I draw on insights from research on teacher cognition and naturalistic decision-making research to identify how perceptions of organizational factors influence instructional decision-making. Results indicate that respondents perceive structural and socio-cultural factors as constraining and affording practice by exerting normative pressures on teaching decisions, imposing logistical constraints, and encouraging autonomy. The relationship between these factors and teaching practice was moderated by factors such as status and other individual characteristics.

Detailed accounts of teaching can shed light on the nature and prevalence of active learning, yet common approaches reduce teaching to unidimensional descriptors or binary categorizations. In this paper, I use the instructional systems‐of‐practice framework and the Teaching Dimensions Observation Protocol (TDOP) to advance an approach to thinking about teaching in science classrooms in more multidimensional terms. Using descriptive statistics and social network analysis, I examine the teaching practices employed by a group of science and engineering faculty (n = 56). Results indicate the extensive use of lecturing with premade visuals (observed in 65% of all 2‐minute intervals comprising a class). However, the majority of instructors (n = 34) lectured for periods of 20 minutes or less. Using the Differentiated Overt Learning Activities (Chi & Wylie, 2014) framework to interpret TDOP codes, the data reveal lower rates of active learning modalities including “being active” (students answering questions, 28%; students problem solving (PS),15%), “being constructive” (students asking questions, 4%; students doing creative tasks, 2%), and “being interactive” (students working with peers to do creative tasks, 2%). Results indicate variation across disciplines and course contexts, that active learning is embedded within PowerPoint lectures, and that small group work exercises are not synonymous with constructivist activities. Implications for research, practice, and policy are discussed.