Pedagogical Knowledge Among Early-Career Undergraduate Instructors: Qualitative Variation, Longitudinal Development, and Influences on Teaching.

Pedagogical Knowledge Among Early-Career Undergraduate Instructors: Qualitative Variation, Longitudinal Development, and Influences on Teaching - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Search Log in Primary site navigation Search Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice CBE Life Sci Educ . 2026 Summer 1;25(2):ar11. doi: 10.1187/cbe.25-09-0196 Pedagogical Knowledge Among Early-Career Undergraduate Instructors: Qualitative Variation, Longitudinal Development, and Influences on Teaching Alex H Waugh † , Kathryn E Green ‡ , Tessa C Andrews †, * Editor: Erika G Offerdahl, Author information Article notes Copyright and License information † University of Georgia, Department of Genetics, Athens, GA, 30602 ‡ Cedar Shoals High School, Athens, GA, 30605 Author contributions: A.W. and K.G. performed the experiments. A.W., K.G., and T.A. analyzed the data. A.W. and T.A. drafted the article. A.W. prepared the digital images. T.A. conceived and designed the experiments. Conflcts of interest: The authors declare no competing financial interest. ✉ *Address correspondence to: Tessa C. Andrews ( tandrews@uga.edu ). Roles Erika G Offerdahl, : Monitoring Editor Received 2025 Sep 5; Revised 2026 Jan 23; Accepted 2026 Feb 4. © 2026 Waugh et al. CBE—Life Sciences Education © 2026 The American Society for Cell Biology. This article is distributed by The American Society for Cell Biology under license from the author(s). It is available to the public under an Attribution–Noncommercial–Share Alike 3.0 Unported Creative Commons License ( http://creativecommons.org/licenses/by-nc-sa/3.0 ). “ASCB®” and “The American Society for Cell Biology®” are registered trademarks of The American Society for Cell Biology. PMC Copyright notice PMCID: PMC13008270  PMID: 41838608 Abstract Active learning can enhance student outcomes in STEM higher education, but its effectiveness varies with implementation. A key contributor to this variation is the pedagogical knowledge held by instructors. However, little is known about instructors’ pedagogical knowledge of how people learn, how these ideas develop over time, and how knowledge development influences active-learning implementation. This longitudinal qualitative study examined variation, development, and instructional implications of pedagogical knowledge among 11 early-career undergraduate life sciences instructors in the context of their active-learning instruction. We conducted semistructured interviews, including stimulated recall, capturing pedagogical knowledge used to plan, implement, and reflect on a lesson, repeating this process across multiple semesters. We used qualitative content analysis and an analytical framework to identify distinct pedagogical ideas about how people learn used by instructors and their alignment with passive, active, and generative cognitive engagement in the ICAP framework. Longitudinal comparisons revealed that participants did not consistently develop ideas aligned with generative cognitive engagement, indicating that teaching experience is necessary but insufficient to foster development of crucial pedagogical knowledge for effective active learning. Case studies illustrated how knowledge development can influence nuances of active-learning design and implementation. We discuss potential mechanisms of knowledge development and instructional implications. INTRODUCTION Active-learning instruction has the potential to increase students’ abilities to learn fundamental concepts, foster the development of scientific thinking skills, and begin to repay educational debt owed to students from historically and currently marginalized groups ( Haak et al. , 2011 ; Freeman et al. , 2014 ; Eddy and Hogan, 2014 ; Ballen et al. , 2017 ; Theobald et al. , 2020 ). Active-learning instruction refers to periods of class time in which the instructor stops lecturing and students work independently or in groups, potentially including metacognition and practice with core competencies ( Collins and O'Brien, 2003 ; Driessen et al. , 2020 ). Importantly, although active-learning instruction can be highly effective at improving student outcomes in STEM, it encompasses a wide range of strategies and each can be used more and less effectively, depending on the nuances of implementation ( Turpen and Finkelstein, 2009 ; Henderson and Dancy, 2009 ; Chase et al. , 2013 ; Daubenmire et al. , 2015 ; Dancy et al. , 2016 ). Given this variation, it is not surprising that the learning gains that students achieve in active-learning classrooms can vary substantially ( Pollock and Finkelstein, 2008 ; Andrews et al. , 2011 ; Nehm et al. , 2022 ). Thus, realizing the benefits of active learning for students depends on convincing instructors to adopt these strategies and supporting them in using strategies effectively. Opportunities to develop key knowledge and skills are one form of support that instructors need (e.g., Stains and Vickrey, 2017 ; Offerdahl et al. , 2018 ; Jackson et al. , 2022 ). Importantly, the knowledge needed for effective active-learning instruction differs from that required for didactic instruction. Rather than focusing primarily on how to repackage and present information so that students can receive transmitted knowledge from the teacher, active learning involves supporting students in constructing their own understanding of the content via generative reasoning that goes beyond the presented information. Creating and sustaining opportunities for this kind of cognitive work may require active-learning instructors to develop and formalize a working understanding of how people learn, a notion underscored by associations between more developed teacher knowledge and more effective implementations of active learning in undergraduate science ( Auerbach and Andrews, 2018 ; Auerbach et al. , 2018 ; Andrews et al. , 2019 ). Pedagogical knowledge is a professional knowledge base for teaching that encompasses ideas about teaching and learning that are generalizable across topic and potentially discipline (e.g., Gess-Newsome, 2015 ; Auerbach and Andrews, 2018 ). The generalizability of pedagogical knowledge makes it distinct from pedagogical content knowledge, which instructors use to teach specific topics to students at a specific educational level ( Gess-Newsome, 2015 ; Auerbach and Andrews, 2018 ; Carlson et al. , 2019 ). In other words, pedagogical knowledge is useful to an instructor across many topics they teach, whereas they will rely on unique pedagogical content knowledge for each topic they teach ( Chan and Hume, 2019 ). Both types of knowledge are likely important for active-learning instruction, and they may interact (e.g., Auerbach et al. , 2018 ; Waugh et al. , 2025 ). Instructors may simultaneously draw on pedagogical content knowledge and pedagogical knowledge to plan and implement active-learning strategies (e.g., Andrews et al. , 2019 ; Waugh et al. , 2025 ). Furthermore, pedagogical knowledge may facilitate the development of pedagogical content knowledge (e.g., Chan and Yung, 2018 ; Andrews et al. , 2019 ; Waugh et al. , 2025 ). Pedagogical knowledge has been the subject of considerably less research than other types of teaching knowledge, so the field lacks clear consensus on its components. Within K12 education, it has been described as including knowledge of how people learn, assessment, lesson structure and logistics, student motivation, approaches to instruction (e.g., Grossman and Richert, 1988 ; Morine-Dershimer and Kent, 1999 ; König et al. , 2014 ). An empirically derived framework of pedagogical knowledge used by undergraduate biology faculty placed knowledge of how people learn 1 as the hub in the middle of six other pedagogical knowledge components because it was the most used and instructors linked it to every other component of pedagogical knowledge ( Auerbach and Andrews, 2018 ). This framework defined pedagogical knowledge of how people learn as an instructor's ideas about how students learn from constructing their own knowledge, engaging in challenging, higher-level work in class, and participating in the practices of sciences, as well as the ways in which instructors create and maintain these opportunities for students ( Auerbach and Andrews, 2018 ). In this study, we focus on pedagogical knowledge of how people learn and broaden the definition to include the full spectrum of instructor's ideas and reasoning about the experiences that foster learning for students, not just those involving generative work. We focus on pedagogical knowledge of how people learn because evidence suggests it may play an especially important role in the implementation of active learning in undergraduate science. Expert active-learning instructors paid more attention than novices to whether instructional strategies provided students with opportunities to construct their own understanding ( Auerbach et al. , 2018 ). Similarly, active-learning instructors who implemented generative active-learning strategies, which foster deeper learning, relied on the knowledge that students learn by explaining their reasoning ( Andrews et al. , 2019 ). In contrast, instructors who primarily engaged students in remembering facts and algorithmic problem-solving did so based on the pedagogical knowledge that active learning breaks up lectures and provides students with exam practice ( Andrews et al. , 2019 ). Given the potentially important role of pedagogical knowledge in evidence-based, active-learning instruction, the development of this teaching knowledge warrants further investigation ( Andrews et al. , 2022 ). Previous research on teacher thinking and expertise sheds some light on the processes of development among college faculty. Adopting new instructional practices, especially with support, may facilitate knowledge development. Engineering instructors who adopted conceptual questions in the supportive setting of a community of practice developed new pedagogical knowledge of how people learn and pedagogical content knowledge ( Koretsky et al. , 2024 ). Some form of intervention or change may be necessary to foster knowledge development for many faculty. Though faculty draw on a range of sources to inform their ideas about teaching and learning, including their own experiences as instructors, students, and researchers ( Oleson and Hora, 2014 ), their ideas may not change much because of teaching experience alone (e.g., Popova et al. , 2021 ). Additional research is needed to better understand the developmental pathways of pedagogical knowledge. Since pedagogical knowledge has only rarely been studied, we drew on additional theory to operationalize pedagogical knowledge of how people learn and serve as an analytical framework. We used the ICAP framework, from cognitive science, which defines active learning by describing four modes of student cognitive engagement: Interactive, Constructive, Active, and Passive ( Table 1 ; Chi and Wylie, 2014 ). This framework links overt student behaviors, as indicators of modes of cognitive engagement, to learning effectiveness ( Chi, 2009 ; Menekse et al. , 2013 ; Chi and Wylie, 2014 ). We used the ICAP framework in a novel way by applying it as an analytical framework to interpret what instructors know about student learning and how they enact that knowledge in their active-learning instruction. This repurposing provided an empirical basis for distinguishing variation in instructors’ pedagogical knowledge and characterizing instructors’ developmental shifts from pedagogical ideas aligned with surface-level cognitive engagement toward those supporting deeper learning processes. TABLE 1: Four levels of cognitive engagement in ICAP framework, described by observable student behavior and expected learning outcomes (adapted from Chi and Wylie, 2014 ). Mode Student behavior Learning outcomes Interactive Two or more learners discuss and substantively respond to one another to generate outputs that go beyond the information that has been presented in instructional materials (e.g., defending and arguing a position) Deepest understanding, potential to innovate new ideas, interpretations, products. Constructive Learners individually generate outputs that go beyond the information that has been presented in instructional materials (e.g., design an experiment, predict the effect of mutations in a pathway) Deep understanding, potential for transfer to new contexts Active Learners recall information and/or make physical manipulations without adding new knowledge (e.g., taking verbatim notes) Shallow understanding, potential for transfer to very similar contexts Passive Learners receive information (e.g., listening) Minimal understanding, potential for knowledge recalled verbatim and in identical context Open in a new tab The ICAP framework has been repeatedly empirically tested across a variety of educational contexts, resulting in robust evidence that students achieve progressively deeper learning and progressively greater potential for transferring knowledge to new contexts at higher levels of cognitive engagement (Interactive > Constructive > Active > Passive; e.g., Chi, 2009 ; Menekse et al. , 2013 ; Chi and Wylie, 2014 ; Freeman et al. , 2014 ; Chi et al. , 2017 ; Menekse and Chi, 2019 ). Students who only engage passively by listening learn less than those who actively engage, such as by taking notes (e.g., Peper and Mayer, 1986 ). Furthermore, students who have opportunities during class to construct something beyond the presented information achieve even greater learning outcomes, and the greatest learning outcomes occur when students have opportunities to engage interactively as they collaboratively exchange and build on ideas with others (e.g., Hausmann et al. , 2008 ; Kramarski and Dudai, 2009 ). The constructive and interactive levels of ICAP are collectively referred to as generative cognitive engagement because both require students to generate something beyond the information presented to them ( Chi and Wylie, 2014 ). If college STEM courses aim to foster deep conceptual learning, the ICAP framework predicts that they will need to include opportunities for generative cognitive engagement alongside more traditional opportunities for passive and active engagement. In this study, we aimed to answer two research questions. First, we asked: what ideas do early-career college science instructors who use active learning have about how people learn? To answer this question, we characterized participants’ pedagogical knowledge of how people learn that they used while planning, implementing, and reflecting on their instructional practices and examined its alignment with the ICAP framework. Second, we asked: what are possible trajectories of pedagogical knowledge development over time, and how does this development relate to instructional practices? To answer this question, we leveraged our qualitative data to illuminate avenues of knowledge development for all participants and examined nuances of how knowledge shifts related to teaching practices using case studies. This study contributes detailed characterizations of pedagogical knowledge of how people learn among life science faculty and illustrates how this knowledge can develop and influence fine-scale instructional practices. METHODS This work was determined by the Institutional Review Board at the University of Georgia to be exempt (PROJECT00000297). Participants We recruited early-career instructors who had taught college courses for four or fewer semesters because we expected this population to be poised to develop their knowledge and practice as they encountered their first teaching experiences as faculty. Given their lack of experience, these faculty may also be especially likely to seek teaching professional development, making our research findings more directly applicable to educational developers. Participants taught large undergraduate life sciences courses (with 50+ students), self-identified as using active-learning strategies, and expressed a commitment to further developing these strategies over time. We systematically analyzed classroom videos (see Analyzing Teaching Strategies) to confirm that each participant stopped lecturing and engaged students in problems and questions for part of class. In total, we collected data from 11 participants spanning 40 semesters of instruction, with interviews and in-class recordings focused on one focal lesson per participant each semester. These participants taught at four public, research-intensive institutions in the Southeastern United States. Each participant held a PhD in a discipline relevant to their teaching. Participants held either a long-term tenure-track ( n = 4) or teaching-track ( n = 7) faculty position. In the United States, where this study occurred, tenure-track positions have research responsibilities and are eligible to earn tenure, an indefinite academic appointment that can only be terminated for cause or under extraordinary circumstance. Teaching-track positions typically do not have research responsibilities and are not eligible to earn tenure. Participants taught a range of upper- and lower-division life science courses, including biology, microbiology, developmental biology, evolutionary biology, anatomy and physiology, and neuroscience, catering to both majors and mixed groups of majors and non-majors. The focal courses had an average enrollment of 135 students (SD = 64.5; Table 2 ). TABLE 2: Participant information. Percentage teaching appointment indicates the percentage of their job responsibilities that were teaching and was reported by participants. Pseudonym Number of time points Course level a Approx. class size Position (Percentage Teaching appointment) Semesters teaching experience as faculty at first time point Amy 7 LD 180 Lecturer (100%) 4 Henry 5 LD 200 Lecturer (100%) 4 Beth 4 UD 80 Assistant professor (25%) 0 Claire 4 LD 100 Lecturer (60%) 0 Dana 4 UD 110 Assistant professor (30%) 2 Eric 4 LD 60 Academic professional (49%) 2 Fiona 3 LD 75 Academic professional (49%) 1 George 3 LD 250 Lecturer (85%) 2 Irene 2 UD 200 Assistant professor (40%) 2 June 2 UD 150 Lecturer (95%) 2 Kim 2 LD 75 Assistant professor (33%) 0 Open in a new tab a UD = upper division; LD = lower division. We collected data from each participant related to a specific lesson in the focal course, repeating the data collection each semester they taught the lesson over two or more semesters (range = 2–7; mean = 3.5; Table 2 ). We collaborated with each participant to select a target lesson that they enjoyed teaching, anticipated continuing to refine, and felt was representative of their instructional approach. For two participants, we captured data related to two different lessons across the study. Amy changed lessons after the first two iterations, and for Henry, we recorded a different lesson during his middle three semesters than in his first and last semester, which were the same lesson. However, for the other nine participants, data came from one lesson. Data collection occurred each semester that participants taught the focal course, once or twice annually. Data collection spanned Fall 2019 to Spring 2023, excluding summers, with the length of participation varying based on when individuals joined the study and when they stopped teaching the focal course ( Table 2 ). Due to the COVID-19 pandemic, some semesters of data collection involved remote teaching or HyFlex formats involving a combination of in-person and remote instruction (Supplemental Figure S1). Though we did not select based on this criterium, participants stood out from typical life sciences faculty because of their extensive involvement in teaching professional development, even in their early-career stages. At the study's conclusion, participants completed a survey about their experiences with teaching development, both before and during the study. The majority (8 out of 11) had participated in over 40 h of teaching development and received formal teaching training during their time as graduate students or postdoctoral researchers. About half (6 out of 11) had engaged in formal teaching mentorships as mentees, and a few (3 out of 11) had even led professional development sessions themselves. Additionally, three participants had published peer-reviewed research on education, and two others had published biology course materials or lessons. Throughout the study, we also asked participants about any teaching professional development they had experienced since the most recent data collection. During their participation in the study, participants engaged in a range of opportunities, such as teaching conferences, formal teaching fellowships that offered peer-to-peer and professional support, bootcamps to support transition to online teaching, and workshops dedicated to improving classroom diversity, equity, and inclusion. Participants’ high level of engagement in teaching professional development likely equipped them with unique knowledge and skills, and suggests high motivation to improve as teachers, which may set them apart from other life sciences faculty. Eliciting pedagogical knowledge of how people learn At each time point, we conducted pre- and post-lesson interviews to capture participants’ pedagogical knowledge as they planned, taught, and reflected on a target lesson, ( Alonzo and Kim, 2016 ; Figure 1A and B ). Interviews also elicited other types of instructor knowledge, which are not the focus of this study. We aimed to characterize practical teaching knowledge closely intertwined with an instructor's actual active-learning instruction, rather than more general beliefs or philosophies. The pre-lesson interview was semistructured and targeted knowledge used in planning a lesson, such as the design of activities and intended implementation. Before the interview, we collected relevant lesson materials such as slides, handouts, and pre-class assignments. Interview questions aimed to explore participants’ pedagogical knowledge, asking how specific instructional strategies in the target lesson supported student learning (refer to full interview protocols in Supplemental Materials, Appendices A and B). After the first semester, interviews began with questions about the reasoning behind changes to the design of teaching materials since the previous iteration of the lesson. Pre-lesson interviews occurred one business day before the lesson and typically lasted around 60 min. FIGURE 1: Open in a new tab Data collection and analysis schematic. (A) Data collected via pre- and post-lesson interviews conducted around a video-recorded target lesson taught by the participant, repeated for 2–7 semesters. (B) Raw data included interview transcripts for qualitative content analysis and videos for systematic characterization of teaching practices. (C) Quantification of alignment with ICAP framework enabled comparisons across participants and time. The quantification system assigned more weight to generative ideas than active ideas. A score of four indicates the presence of three generative ideas: To explain is to learn , Application and practice , and Generative work is a mirror . A score of three indicates the presence of two of these three ideas. A score of two indicates the presence of one generative idea: Application and practice or To explain is to learn . A score of one indicates only active ideas ( Breaking up lecture , Peer-to-peer listening , Struggle makes it stick ). Interviews with Refrain from explaining increase score by one (refer to Supplementary Table S2 for the full quantification system). Scores from the pre and post were summed to produce one score per time point per participant, which we plotted over time to observe patterns of change. The design of post-lesson interviews incorporated opportunities for participants to view short video-clips of their own teaching from the target lesson to stimulate their ability to recall their thinking during instruction ( Alonzo and Kim, 2016 ). To generate these clips, we video-recorded each target lesson using a camera and a lapel microphone worn by the instructor. These recordings also served as a source of raw data for a systematic analysis of teaching strategies (see Analyzing teaching strategies). Before each post-lesson interview, we selected three or more clips (range = 3–6 clips) from the recorded lesson. Filmed from the back of the room to better preserve students’ anonymity, clips showed students working and, when possible, instructors interacting with students as they worked. Following a clip selection protocol (Appendix B), we prioritized lesson segments that showed instructors interacting with students as they engaged in generative cognitive work. When present, we also included contrasting lesson segments that showed students engaged in “active” cognitive work, which involved recall or algorithmic problem-solving. Post-lesson interviews occurred one to two business days after the lesson and typically lasted around 60 min. During each post-lesson interview, participants and the interviewer watched selected video clips from the recorded lesson. We asked participants to provide a “running commentary” of what they were thinking during the moments shown in the clips. After they shared their recollections, we asked follow-up questions to delve more deeply into their thought processes. Since the content of the clips varied based on the lesson, the interview protocol had to be flexible. For each interview, we selected a customized set of questions drawn from a larger pool. We asked each question in the larger pool at least once during each interview, ensuring that interviews were consistent across time and participants, despite the variation in questions from clip to clip. Questions included, “In this moment, what were you thinking about your role in this interaction with the student? Why is that role important?” and “How does this [general instructional strategy, such as think-pair-share, case study, or worksheet activity] support student learning?” (refer to full interview protocol in Supplemental Materials, Appendices A and B). Qualitative content analysis of pedagogical knowledge We aimed to identify and characterize pedagogical ideas of how people learn that participants used when planning, teaching, and reflecting on a target lesson. We developed codes both deductively and inductively. We drew on prior studies of pedagogical knowledge among undergraduate instructors and the ICAP framework to develop a priori codes to capture pedagogical knowledge of how people learn (e.g., Saldaña, 2003 ; Charmaz, 2006 ; Auerbach et al. , 2018 ; Auerbach and Andrews, 2018 ; Andrews et al. , 2019 ). Variation in the data led us to create emergent codes within and beyond a priori codes. We used ICAP as an analytical framework to guide our sensemaking of this variation, resulting in the final set of seven pedagogical ideas about how people learn each aligning with either the generative, active, or passive mode of cognitive engagement. Generative ideas about how people learn to include language that implied learning occurs when students go beyond presented information as they (1) apply their knowledge in novel scenarios, (2) practice using knowledge and skills, or (3) generate externalized representations of their thinking (e.g., verbal/written explanations). Active pedagogical ideas connected the process of learning to the work that goes on during in-class activities, but these ideas did not specify that this work should be generative in nature, involve practicing skills, or require students to apply their knowledge in novel scenarios. Passive pedagogical ideas implied that learning occurs through listening or were otherwise vague as to the mechanism of learning. At least two researchers (AHW, KEG, and TCA) independently read and coded each transcript, and any disagreements were discussed until a consensus was reached. As we refined the definitions and boundaries of the codes, we reanalyzed all previously coded data to ensure consistency. The result of this process was a finalized codebook (Supplemental Materials, Appendix C). We developed this codebook during the early stages of the project, using about 50% of the data, and then applied it to the full set of interviews over time for each participant. Next, we aimed to assess the extent to which participants’ pedagogical knowledge of how people learn developed in alignment with the ICAP framework (i.e., from a focus on active to generative). This required reducing our rich qualitative data to a form that could be compared across time. We developed a system to quantify the presence of certain ideas about how people learn and combinations of ideas ( Figure 1C ). Guided by the ICAP framework, we assigned higher scores to generative ideas about how people learn (refer to Supplementary Table S2). We separately quantified participants’ knowledge demonstrated in the pre- and post-interview because some participants used ideas in planning that they did not later rely on when implementing a lesson, whereas others consistently applied the same ideas in planning and teaching. We summed the pre- and post-interview scores for each time point and plotted scores for each participant over time to identify patterns in knowledge development toward ideas aligned with generative cognitive engagement. Case study selection and analysis Though quantifying participants’ pedagogical knowledge illuminated trajectories over time, it also necessarily reduced interesting and important nuances in participants’ development. Therefore, to take full advantage of the rich details present in our dataset, we next developed case studies to characterize the details of how instructors developed pedagogical knowledge of how people learn and how knowledge development corresponded to changes in teaching practices. Given the limited research on pedagogical knowledge development in college faculty, we used a case study approach to illuminate the nuanced and context-dependent ways this development unfolds. As such, we used a holistic approach to select participants for cases who (1) exhibited a substantial (3+ point) increase in their knowledge score in at least one semester and (2) developed and maintained new pedagogical ideas corresponding to the generative levels of the ICAP framework. These selection criteria left us with three options for case studies: Beth, Dana, and Irene. For each of these three, we began by compiling their teaching materials, video-recorded lessons, and all evidence of their knowledge of how people learn from each time point. We summarized the participants’ knowledge at individual time points by identifying the ideas present through the applied codes and capturing nuances that extended beyond what the codes represented. We also documented which teaching materials had changed, and when, as well as whether and how implementation had changed. Coarse-grained observation protocol data (see Analyzing Teaching Strategies) added to our ability to assess teaching practice changes. After compiling evidence of pedagogical knowledge development for each of these participants and writing initial descriptions, we selected Dana and Irene as case study subjects based on the overall clarity of the participants’ quotations for the purposes of reporting. We continued to refine our case descriptions of Dana and Irene's knowledge and practice development until two researchers (AHW and TCA) agreed that the case description was complete, accurate, and accessible to a wide audience. Analyzing teaching strategies To answer our second research question, we drew on insights from our qualitative findings to examine how participants’ developing pedagogical knowledge influenced their teaching practices. This required a systematic analysis of participants’ teaching practices for each target lesson. We conducted this analysis, focusing on three specific behaviors: instructors interacting with students in small groups or one-on-one, students working individually or in small groups during class, and students explaining their reasoning to the instructor. The first two behaviors were adapted from the Classroom Observation Protocol for Undergraduate STEM ( Smith et al. , 2013 ), and we developed the third to capture instances when the instructor heard student reasoning. We defined students explaining their reasoning as any instance where a student verbally elaborated on their thinking to the instructor. This could happen in various formats, one-on-one, in small groups, or in front of the entire class, as long as the instructor appeared to be listening. To ensure we captured elaborations, we only counted explanations that involved more than a single word or phrase. To quantify these behaviors, we coded each 2-min segment of the video-recorded classes, marking the presence or absence of each behavior. Because the behaviors are not mutually exclusive, we coded each one independently, meaning a segment could receive any combination of the three behavior codes. We then calculated the average percentage of segments containing each behavior across the target lessons for the instructors featured in case studies. A coding team, trained as part of a larger project, followed a rigorous protocol to ensure consistent analysis. To assess inter-rater reliability (IRR), we used Cohen's kappa, which ranges from 0 to 1, with 0 indicating no agreement above chance and 1 indicating perfect agreement after chance correction ( Cohen, 1960 ; Landis and Koch, 1977 ). Higher κ values reflect greater consistency in how coders applied the coding scheme. Once high IRR was achieved, individual coders analyzed the remaining videos, recalculating IRR periodically to maintain consistency over time. Of the 40 videos analyzed for this study, 16 were analyzed by multiple coders, resulting in an average Cohen's kappa of 0.96, indicating strong agreement among coders. We recruited participants based on their self-reported use of active-learning instructional methods. Using our quantification of the student working behavior, we confirmed that each participant stopped lecturing and engaged students in problems and questions for part of class. When focal lessons were held in-person with no remote option, students worked for an average of 43% ( N = 23 in-person semesters, range = 15%–100%) of 2-min lesson segments. A summary of behavioral data from these semesters is included in Supplemental Table S1. Trustworthiness of these qualitative analytic approaches Trustworthiness in qualitative research is established through four key domains: confirmability, dependability, credibility, and transferability ( Guba, 1981 ; Anfara et al. , 2002 ; Shenton, 2004 ). In this study, we incorporated specific practices throughout data collection, analysis, and reporting to ensure a trustworthy approach to our qualitative work. Confirmability and dependability are critical in maintaining the integrity of qualitative data interpretation. Confirmability refers to ensuring that interpretations accurately reflect the original data, while dependability focuses on the consistency and repeatability of the findings by different researchers ( Anfara et al. , 2002 ; Shenton, 2004 ). To achieve this, we employed constant comparison methods, with multiple researchers deeply engaging with the data. By requiring all interpretations to be discussed collaboratively and reach consensus, we mitigated the risk of individual biases influencing the outcomes. This process ensured that our analysis was not driven by any single researcher, enhancing both confirmability and dependability. Credibility, another domain of trustworthiness, refers to the alignment between the research focus and the methodology, ensuring that the study measures what it intends to ( Shenton, 2004 ). Our data collection methods were grounded in established literature on pedagogical knowledge, using well-documented approaches such as stimulated recall ( Alonzo and Kim, 2016 ). These methods were specifically designed to explore early-career instructors’ pedagogical knowledge in planning, enacting, and reflecting on instruction, ensuring that the data accurately addressed our research questions ( Andrews et al. , 2019 ). Lastly, transferability involves assessing the extent to which the findings can be applied to other contexts ( Shenton, 2004 ). Although we do not claim generalizability given our sample size, participant diversity, and the context-specific nature of knowledge development, we support the transferability of our work by providing rich descriptions of our participants, their teaching environments, and their institutional settings. This level of transparency allows others to assess the relevance of our findings to similar contexts. Additionally, we use precise language to limit our conclusions to the study population, acknowledging the specific context and scope of our research. RESULTS We present results in three sections. To answer our first research question, we characterize qualitative variation in participants’ pedagogical knowledge of how people learn exhibited in pre- and post-lesson interviews and how this variation aligned with the ICAP framework. This led to the identification of seven recurring ideas about how people learn that aligned with the passive, active, or generative levels of the ICAP framework. To answer our second research question, we first present longitudinal quantification of pedagogical knowledge development, which revealed trajectories over time and patterns across participants. However, these quantifications also reduce important complexity in how participants’ knowledge developed, so the third section presents in-depth case studies of longitudinal pedagogical knowledge and teaching practice development focused on two participants. Unless otherwise indicated, all text within quotation marks represents verbatim utterances from participants, with minor editing for grammar and clarity. All participant names have been replaced with pseudonyms. Qualitative variation in pedagogical knowledge of how people learn and alignment with ICAP Participants relied on a range of ideas about how people learn as they planned, enacted, and reflected on active-learning lessons. The next sections describe seven distinct pedagogical ideas about how people learn starting with ideas more aligned with the passive and active ends of the ICAP framework and proceeding to ideas about generative cognitive engagement ( Table 3 ). Participants shared most of these pedagogical ideas as the basis for their teaching strategies. Two ideas, however, stood out from the rest. One, which we termed Generative work is a mirror , dealt with the mechanism by which participants thought generative work fostered student learning. Participants used another pedagogical idea, which we refer to as Refraining from explaining , to describe their intentions as they facilitated student work in real-time during class. TABLE 3: Pedagogical knowledge of how people learn that participants relied on as they planned, enacted, and reflected on their active-learning instruction, with example quotes. Pedagogical ideas about how people learn Exemplars Breaking up lecture: Learning occurs when students have opportunities to engage actively with tasks in the classroom rather than passively listening to lecture ( Active ) “If you're just listening, you're passively learning. If you're writing or talking or gesturing, I think those are signs of probably more engaged learning… I think those activities help them become more active in their learning.” – Beth “The fastest way to prepare for me is just traditional, boring lecture using previous PowerPoints that other people have designed. But I find that the students are not asking questions, and they're really confused [and] aren't able to apply that content to the next lecture. And so, when I started adding in these break points to the lecture to do little activities or get feedback from them on and talking to their neighbor and reporting back, I found that they were more engaged.” – June Peer-to-peer listening: learning occurs when students hear ideas from peers ( Passive ) “I suppose it's almost similar to a think-pair-share exercise in terms of that they then get to talk it through. They get to hear other people's way of thinking about this. It gives them an opportunity to update their thinking based on what their neighbors are thinking.” – Irene “So, I think my hope with the groups was that it would provide some knowledge sharing between the students.” – Kim Struggle makes it stick: learning occurs when students remember what they struggled with during a difficult task ( Active ) “I feel like if you struggle with something and get it wrong, then you might be more likely to kind of work through that instead of I just tell them this and then later they're like, ‘How do I remember? I don't know.’” – Beth “I think my goal in designing it is— I do think that when you're asked to think about something hard first, before you get any information…that then it sticks. Like you've created a framework or context that then you can slot that new information in.” – Kim Application and practice: learning occurs when students practice skills and apply their knowledge to novel scenarios ( Generative ) “You get a little immediate practice because, too often, I think we explain something in class, class ends and then we say, go do your homework. No one does their homework the first day. You know, I never did my homework like immediately, and by the time you come back to it, you've forgotten most of it. Using the clicker gives them a little bit more review to kind of reemphasize, ‘Here's the term we went over, it's important. I want you to get out the pencil and paper and do the math, you know, kind of build that muscle memory’.” – George “This class is very application focused. It's all about applying information, not just knowing it. So, um, if you approach the class from, you know, reading the textbook and just watching videos, or just going to class. That, because I know this information that I understand it, that's where the struggles come because you don't really understand it. You're very familiar with it, but you don't really understand it because you can't apply it.” – Henry To explain is to learn: learning occurs when students outwardly express their thinking, often via verbal and/or written explanations ( Generative ) “I always think about [student-led learning] in terms of anything that the students are doing, where their voice is the one sharing the knowledge. In this case, they're sharing… It's giving them the opportunity to explain. I feel like every time that I truly know something is when I can explain it well and someone else understands. Giving them that opportunity to do that might help them realize that they also know it by explaining it.” – Fiona “I prefer that they work on them together with other students because it generates more outward expression of what you are thinking. So, the students, rather than just being an internal dialogue with yourself, as far as what you're doing, when they're working in groups, they have to explain to each other what they're thinking, bounce ideas off of each other.” – Henry Generative work is a mirror: learning occurs when students reflect on their own thinking because of engaging in generative work ( Generative ) “… sometimes our students don't know what they don't know and they're really clever. They're really bright. If they feel like they recognize things, they tend to think they know more than they did. But when we give them this opportunity that's just a little bit more challenging and they continue to be challenged by it, that might be that realization, ‘I actually don't know this as much as I thought I did.’” – Fiona “Instead of just listening and digesting, now they're applying and making predictions… And then, if they predicted wrong, then maybe they can think about, ‘Okay, what was my line of thinking here and why was that wrong?’ So, it's getting them to be a little bit more actively participating with what's being presented to them” – Claire Refraining from explaining: learning occurs when instructors keep the onus for generative work on students by prompting students to articulate their reasoning before explaining answers ( Generative ) “Let the students drive it, right? I know what the answer is, and I'll show the answers that I was thinking of in a bit, but I don't want to constrict them too much from the start because when they come to the conclusion on their own, then it's more cemented… So, trying to be as hands off and more of a guide in the right direction.” – Eric “I'm trying to identify how far through this thought process or through this skill they're able to navigate on their own and then see how little I can give them to go the rest of the way…If they just need a tiny little push or a tiny little point in a direction, that's making them use a lot more brain power and a lot more of these skills we're trying to build than if I kind of told them they were wrong and then told them the right answer. So, giving them as little as possible to just kind of nudge them down the road, get them started again is what I hopefully try to do all the time.” – Amy Open in a new tab Breaking up lecture keeps students engaged and creates opportunities to go beyond recall. Since all participants used active-learning strategies in their course, it was not surprising that they saw advantages of active learning over traditional lectures. Participants explained that they used active-learning strategies because they created opportunities to break up lectures with active tasks, which they thought helped sustain students’ attention. For instance, Eric asked a discussion question near the beginning of class and explained the advantage of it this way, “I think I've talked for about 5 or 10 min here. This [question] is just to get them engaged, right? To break up the time of me talking and droning on.” Participants commonly expressed this idea. In some cases, participants who expressed this pedagogical idea seemed to hold an underlying assumption that students learn best by listening to lectures but needed attentional breaks afforded by active-learning tasks. In their view, these breaks would allow students to maximize their learning by enabling them to continue to pay close attention during lecture. Thus, their pedagogical knowledge of how people learn framed active learning as a tool to keep students attentively engaged during lectures, but this simultaneously positioned the passive cognitive engagement (i.e., listening to the lecture) as the most important learning opportunity in class, a view out of line with the ICAP framework ( Chi and Wylie, 2014 ). Participants also used active-learning strategies because they wanted students to do more than memorize information. Kim explained, Maybe this is a little controversial, but I don't see any value in memorizing material. I memorized the Calvin cycle on like four different occasions over the course of my education and promptly forgot it immediately afterwards, until I needed it. And we live in an age where you can look up anything in an instant. So, I no longer really think that the important thing we're doing here is getting them to understand the Calvin cycle. Instead, I think, in an age where you can Google any fact in a minute, what's much more important is to be curious so that you do look up that information and then to know what to do with the information once you have it. (Kim) Similarly, Amy aimed for students to do more than passively receive information. Rather than give students a table of completed information, Amy asked students to fill in the table with their predictions, If I had shown them this table without giving them a chance to fill it in themselves, there would've been no reason for them to think about why this happens or how this actually looks in reality. They just would've copied it down and then memorized it.” (Amy) Here, Amy conveyed a common idea among participants: passive delivery of information to students would not achieve the type of learning she hoped to foster in her course. Peer-to-peer listening allows students to hear ideas from each other. Every participant engaged students in some form of small-group work during class, and some participants valued group work primarily or solely because it created an environment in which students could learn from their peers ( Table 3 ). For instance, Amy explained that students were, … all coming with different strengths as far as biology and that stuff goes. So, by giving them a chance to talk to each other, the ones that have more advanced prior knowledge are able to help those other folks along. (Amy) Here, Amy is describing how group work fosters learning because students can learn from more knowledgeable peers. However, it is unclear if the cognitive engagement of the students she believes have less advanced prior knowledge goes beyond passively listening. Some participants thought students’ peers could offer different and potentially more relatable explanations than the instructor. When asked to explain how talking to each other helps students learn, George responded, I think they hear a different voice. I think it's that simple. I think hearing it from somebody else who's learning it [for] the first time, they might have a different way of explaining it or thinking about themselves that might resonate with one of them. (George) Overall, this pedagogical idea prioritizes students listening to each other, which is passive engagement. This finding is noteworthy because, although participants described using group work to move away from the passive learning associated with traditional lectures and toward more generative learning, their rationale often mirrored that of lecture-based instruction when they emphasized learning through passive listening to a knowledgeable peer. In later sections, we show that while some participants held only this passive view of group work, others expressed it alongside more generative ideas about the learning benefits of collaborative engagement. Struggle makes it stick because difficulty focuses student attention. At times, participants discussed the relationship between the difficulty of a task and its effectiveness for student learning. When participants displayed this pedagogical idea, they explained that they designed active-learning opportunities to expose students to challenging tasks. For instance, Henry used questions he thought of as more straightforward, as well as more difficult questions, and he explained his reasoning for the latter, I think struggling is good in general in class… It's sort of an unsubstantiated belief that we tend to remember things that we struggle with. Things that come very easily to us we don't remember quite as well for whatever reason. (Henry) Here, Henry explained that the difficult tasks seem to help students remember the content. Typically, the idea Struggle makes it stick lacked an underlying mechanism that explained how difficult tasks support student learning. However, when participants relayed a mechanism, it usually echoed Henry's idea that struggling facilitates later recall. Other participants talked about how they expect challenging tasks to enhance learning by encouraging students to pay more attention to specific topics later when they study, as the difficulties they encountered in class can make those topics more memorable. Interestingly, this perspective positions the learning benefits of active learning as occurring outside of class, as well as in class. As an example, Beth explained, … maybe having some wrong answers in class will inspire some of them to get that information pulled out and together before the test. (Beth) In general, the idea that struggle fosters learning hinted at, but fell short of higher levels of the ICAP framework because difficult questions do not necessarily require generative work. Application and practice foster deeper learning and skill development. Participants often used active learning because they wanted students to apply their knowledge to novel scenarios and practice using skills. Since they expected this level of mastery on summative assessments, they used class time to engage students in this work. Sometimes participants aimed for students to practice using terminology and definitions. Fiona's lesson involved an opportunity for students to practice distinguishing between two related concepts in a real-world scenario, I think it's helpful because, you know, sometimes we just need practice with those vocabulary things… [Students] may feel like they know it in the moment and then, when we're asking them to apply it for the exam, they're not able to because they haven't practiced that way. They haven't practiced that application-type question. (Fiona) Other times, participants explained that certain concepts or skills required repeated exposure and practice to develop fluency, and they used active learning to offer that practice to students. For Dana, in-class activities offered opportunities for students to gain additional experience, in a low-stakes environment, applying their knowledge in a complex, novel scenario, … part of the logic of these assignments is modeling the thinking that really, they should be able to do for assignments and tests and things like that. So, they have a couple of these questions on exams and part of this interaction is you have practice with it. It is not something that is super intuitive. It does take a little bit to get your head around it. (Dana) Thinking beyond exam preparation, some participants viewed opportunities for application and practice as essential for learning. For instance, Fiona explained how chances to apply their knowledge both prepared students for exams and fostered the type of thinking abilities that students need as people and professionals, We've very much moved away from just regurgitation of knowledge on the exams and more into the application. As a practicing professional, you will have the ability to look things up. That is the reality. You won't have the thinking skills unless we really help you develop those. And so that application is for both. Can you apply what you just learned and hopefully remember and take it with you? But, also, practice applying what you learned and thinking through and reasoning out the different scenarios because that's really what the long-term goals for the program are. That's what a lot of people hope to gain after a university education, and so it aligned better. (Fiona) Though participants described different goals for engaging students in knowledge application and practice, from applying concepts to solving problems in novel scenarios, these ideas focused on engaging students in generative work ( Chi and Wylie, 2014 ). To explain is to learn: outward expressions of reasoning foster deeper learning. Participants offered the most robust evidence that they prioritized generative cognitive engagement when they articulated that students learn the most from expressing their reasoning about a topic, typically via verbal explanations or written responses to questions. Kim talked about how the process of explaining one's reasoning during group work is crucial to learning, [Students] can take on a teaching role and that can help their own learning. I think the best way you learn something is by explaining it to somebody else. So, I'm thinking that can help them feel challenged. Like they're being asked to really explain it to somebody else. (Kim) This quote emphasizes that students learn through group work because of opportunities to explain ideas to others, a generative opportunity, which is notably different from the idea that students learn from passively hearing peers’ explanations. Participants also described the learning value of students expressing their reasoning in individual work. For instance, Beth implemented a short writing assignment in the middle of class, and described how this helped students learn, I think when they have to write down an answer to the question, I think that helps them to be focused on a task to completion. You know, if I ask them to think about something, they can think about it, but them thinking about it doesn't really look any different from them planning their grocery list or whatever. When they have to write, they're pretty engaged in the task and they're going to finish their answer and writing is a little slower than thinking. And so, they're going to have to go through a thought process, whether that thought process is correct or not. (Beth) Here, Beth distinguished between the work of outwardly expressing their thinking versus a strictly internal mental exercise, emphasizing that externalizing their reasoning forces students to engage more deeply with the material. Typically, participants who expressed this pedagogical idea wanted to create these opportunities in order to foster deep conceptual learning, but sometimes they described how explaining one's reasoning helped students recall content. As was true for the pedagogical ideas described above, participants did not necessarily share a mechanism for how explaining one's thinking fosters learning, but, as we outline in the next section, some did. Generative work is a mirror that helps students realize what they do and do not know. Among participants who displayed generative pedagogical ideas, some articulated a mechanism for this learning. These participants explained how generative work allowed students to realize what they do and do not know. For instance, Kim contrasted how students tend to evaluate their own understanding favorably when they hear an explanation from someone else but then realize the shortcomings in their thinking when they are forced to explain to someone else, I think it's easy to feel like you understand something when you hear it. Often, you don't realize the things you don't understand until you try to explain it to somebody. I feel like that all the time where I'll start teaching something and I'll realize like, ‘Oh, there's some nuance to this.’ Kim reflected on her own experience recognizing the limitations of her knowledge as an instructor and aimed to leverage those insights to create similar opportunities for her students. As another example, Amy explained that there was learning value in asking students to make a prediction by drawing a graph of anticipated results, I think it's just like concrete…once their pencil hits the paper, even if they think they understand it, just kind of listening to it, they can more easily move on without actually coming to terms with that in their own head. But if I'm asking [them] to actually draw something on their paper, it's in that moment that they're like either they know it or they don't, they're going to draw something or they're not going to be able to. And that's either a wakeup call that they don't necessarily understand what it is I'm asking, or what this concept is, or they do, and they can make that connection. Most commonly participants linked the pedagogical idea Generative work is a mirror to the idea To explain is to learn . However, a few participants linked this mechanism to opportunities for application and practice, explaining that practicing a skill or attempting to apply a concept reveals to students what they do and do not know. Refraining from explaining keeps the onus on students to engage in generative work. In a few instances, participants described their responsibility, as facilitators during class, to maintain the opportunities they had created for students to engage in generative work. They recognized that nuances in how they facilitated in-class work could undermine the cognitive demand required of students during carefully designed generative tasks. Participants described how they had to refrain from offering their own explanations and instead prompt students to explain their reasoning. For example, Beth described how she consciously avoids providing answers to student questions, I almost never give students the answer. I will do something like kind of nudge, nudge, nudge, and then when they get close to the right answer, I won't even necessarily stick around to acknowledge it. I'll say like, ‘I feel like you've really got a handle on this. I'm going to walk away. So, you can finish making this connection yourself and then call me over if you need more help.’ So, I like the idea of them getting there on their own steam, and I know they can do it. (Beth) Beth explained how she will often leave an interaction with students when they are close to arriving at the correct understanding because she wants students to make connections on their own. Participants who used the idea Refraining from explaining often distinguished between two approaches to interacting with students during opportunities for generative work: the instructor explaining the answer and the instructor prompting students to explain their answer. Of note, Dana recognized her natural tendency to explain concepts to students, I've been trying to work on my own tendencies. Especially when I see [a student] has a question, I tend to go into, ‘Okay, let's just go through the answer’ instead of going through prompts to the students, because that definitely takes more time. (Dana) We followed up by asking what was important about prompting students rather than giving information, and Dana responded, It's really just making them verbalize the process… if they are walking me through their logic, it's a way to see where in the process is there a discrepancy. It's very easy if I'm just explaining something, students are like, ‘Oh yeah, this totally makes sense. No problem.’ But when it comes to actually explaining why or doing it on their own, then sometimes a little more challenging. So, it's a way to practice all of that and force that practice. (Dana) Intrigued by Dana's idea that she might need to “force” this kind of cognitive engagement, we asked Dana to elaborate, I know that it is an easy, easier thing to fall into, ‘Well, I'll just let somebody else explain it and just listen to them.’ Sort of passively get that information as compared with, ‘I actively need to work through this on my own’. The idea that, of course things make sense when somebody who knows what they're talking about is explaining it. You can follow along, but can you tell that same logic? Can you repeat that information? That is a different thing. (Dana) Dana expressed that students may tend to passively engage with an activity by letting the instructor explain concepts to them, which meant that she needed to resist her own tendency to explain and be intentional about prompting generative work in one-on-one interactions with students. In other cases, participants described maintaining the cognitive demand of their in-class tasks by withholding certain explanations, in order to leave room for students to independently connect concepts or apply knowledge from another lesson. Henry did this when he intentionally omitted information from his introduction of a problem that students would need to solve the problem. Henry did so to ensure that students had the opportunity to apply to a concept they had already encountered in other lessons to a novel scenario, I never expressly said what direction the anticodon would have to be in during my explanation. And that was just to see if they could figure out on their own that, like in all previous cases, the two have to be in opposite directions. So, I was hoping that they could put that together. (Henry) Henry described that creating conditions for generative work requires thoughtfully designed tasks that pinpoint important conceptual leaps then deliberately launch a task in a way that avoids undermining students’ opportunities to learn by making those conceptual leaps for themselves. Longitudinal quantification of pedagogical knowledge development Quantifying participants’ pedagogical knowledge at each time point reduced the complexity of our rich qualitative data, but, as we report here, it allowed us to detect patterns across time and among participants. Overall, not all participants showed clear patterns of developing pedagogical knowledge that aligned with the generative levels of the ICAP framework over time ( Figure 2 ). Higher scores indicate more knowledge related to generative work. Some participants exhibited a large single-semester increase, a few exhibited a more gradual developmental trajectory, whereas others showed an overall decrease by the end of the study ( Figure 2A ). Participants more commonly increased scores between time points ( n = 14, 48%) than remaining stable ( n = 7, 24%) or decreasing ( n = 8, 28%). Some participants, like Dana, exhibited more stability in their knowledge than others, meaning that once they developed a new idea, they continued to display it in subsequent interviews ( Figure 2B ). Other participants exhibited less stable knowledge development, such as Amy ( Figure 2B ). Seven out of 11 (64%) participants increased their score by more than one point by the end of their participation ( Figure 2A ). This increase in value corresponds to the addition of one or more of these pedagogical ideas: Application and practice , To explain is to learn , Generative work is a mirror , Refraining from explaining ( Table 3 ). Examining individual pedagogical ideas, we found no discernible patterns to suggest that one idea consistently emerged before or alongside another, or that some ideas were more easily lost than others. We also found no discernible patterns to indicate that course level, class size, percent teaching appointment, or teaching experience influenced pedagogical development in our sample. Overall, these data suggest that pedagogical knowledge development is not a guaranteed result of gaining teaching experience, and that instructors are more likely to gain and continue to use generative ideas about how people learn than to cull these ideas. FIGURE 2: Open in a new tab Longitudinal trajectories of pedagogical knowledge of how people learn, by participant. (A) Quantification of pedagogical knowledge, by time point. Scores are sums of knowledge expressed in the pre- and post-lesson interviews. Higher scores correspond to a greater number of generative pedagogical ideas. The minimum score is two, which represents an absence of generative ideas in both interviews. The max score is 10, which represents the presence of all four generative ideas in both interviews ( Table 3 ). A more detailed description of the quantification method is provided in Supplementary Table S2. (B) A tile plot showing the presence of pedagogical ideas by time point and whether participants expressed the idea in one interview (light blue) or both interviews (dark blue). Generative ideas are labeled with purple text, whereas active and passive ideas are in black text (refer to descriptions in Table 3 ). An expanded version of this figure, including lesson modality, distinctions between pre- and post-lesson data, and instructional practices, is provided in Supplementary Figure S2. Case studies of pedagogical knowledge development and instructional practice change We present two exemplar cases of knowledge development, with the goal of uncovering ways in which this knowledge can develop and its potential interplay with teaching practices. While quantifying knowledge helped identify overarching patterns of knowledge development, it offers a limited view compared with the depth and nuance captured in our qualitative data. A case study approach complements our quantitative analysis by elaborating on qualitative knowledge development, revealing how participants generated new ideas and applied them to their teaching. The first case study describes how Irene developed the pedagogical idea To explain is to learn before she linked this idea to Generative work is a mirror . Irene then implemented teaching practices that created more robust and frequent opportunities for this kind of cognitive work in the target lesson. The second describes how Dana developed knowledge about Refraining from explaining and eventually began eliciting student reasoning during the target lesson. Case 1: Irene develops and uses pedagogical knowledge that to explain is to learn as she refines her teaching practices and materials. As part of her upper division evolutionary biology course, Irene taught her target lesson on sources of phenotypic variation (i.e., genes, environments, gene by environment interactions). Safety precautions during the COVID-19 pandemic necessitated virtual, synchronous instruction in semester one and a hybrid format in semester two, with half of students attending in-person and the other half attending remotely. During the target lesson, Irene alternated between lecturing and giving students opportunities to work on active-learning assignments, including open and closed-ended questions (e.g., multiple-choice, true-false). Irene's students worked in roughly 25% of two-minute lesson segments in both semesters. In the first semester's virtual lesson, Irene did not interact with students one-on-one, mostly likely due to the challenges of virtual instruction, and in the second semester's hybrid lesson, she did so in 6% of the 2-min lesson segments (Supplementary Figure S1). In both semesters, Irene did not hear students explaining their reasoning at any point in the lesson. Though these coarse-grained measures of instructional practices did not meaningfully change across Irene's two lessons, Irene made important changes in the design and implementation of her active-learning assignments. As we describe below, these changes aligned with her developing pedagogical knowledge of how people learn via explaining their reasoning. In semester one, Irene relied on ideas about active learning benefiting students through opportunities to practice applying knowledge and activity being better than passivity and did not express the pedagogical idea that students learn by explaining their reasoning. Irene shared that: … being an active participant in the process and applying is just a much better way of learning. They're applying their knowledge and sometimes the questions are simple in terms of—they just have to regurgitate… But most of the time I'm trying to get them to apply. (Irene) Irene explained that a teaching professional development workshop from her time as a postdoctoral researcher had influenced her decision-making on which types of activities to include as active-learning assignments. Irene explained, I suppose the one thing that I always took from the [organization name] workshop was that you can, even if you've got a big class and multiple-choice questions, you can ask those higher learning things you don't just have to [ask basic recall questions]. (Irene) Irene also expressed the idea Generative work is a mirror as she described students interpreting graphs and “think[ing] critically about the topic, I suppose, from things that I've read on my own and personal experience, doing is better than listening and that is where things start to make sense. Or you realize they don't make sense and then you can ask for help. It's to get them to realize if they are understanding or if they're not understanding. (Irene) In the second semester, Irene expressed a new piece of pedagogical knowledge: to explain is to learn . This development in her knowledge coincided with changes in how she implemented active-learning assignments in the target lesson, as well as the questions she posed to students on slides and verbally. During the target lesson in the second semester, she encouraged students to talk in small groups before answering poll questions. Then, after the class saw the results of the initial poll, she encouraged students to again discuss in small groups before polling a second time. In the first semester, she only polled students once and no opportunity for peer discussion followed the poll. In the second semester, Irene told us, This is new for me this year. This was actually suggested to me by our [center focused on teaching and learning]. I really like the suggestion because it allows the students to think about their answer. (Irene) The distribution of students’ responses shifted towards a majority selecting the correct answer after the second poll. Irene explained why she thought this shift had occurred, If you just poll them [once], they're not having to explicitly explain why. You could just pick an answer, but talking to somebody, you actually have to say why… By that, you are having to engage with the material more to have that conversation, I think, and that can lead to a deeper understanding of it and just realizing that maybe, ‘Oh, my initial impression was wrong.’ It gives them time to re-evaluate it. (Irene) In addition to adding the idea that explaining one's reasoning fosters learning, Irene offered that the mechanism leading to learning is students realizing that something about their initial thinking was wrong (i.e., Generative work is a mirror ). Irene also modified the framing of questions she posed to students in ways that created more opportunities for generative work. In semester one, Irene posed a multiple-choice question on a slide that included two possible processes that could result in an observed pattern and an option to indicate that there was “not enough information to distinguish,” which was the most scientifically accurate choice. She then described an experiment that could be conducted to determine which process caused the observed pattern and displayed two graphs with data that could result from the experiment. The text on the slide asked students which graph corresponded with one of the processes and why. However, when Irene verbally posed the question during class, she did not ask for reasoning and told students “If you feel like it, feel free to put this in the chat, and then we'll discuss the answer.” Students submitted answers but did not share reasoning behind their selection. In the second semester, Irene asked an open-ended question rather than a multiple-choice question about which process could drive the observed pattern and did not suggest that insufficient information had been provided. Students discussed with their neighbors and entered responses ranging from one word to a short sentence or two. Irene displayed these responses to the class and pointed out that some students had chosen each process, and that they were collectively correct to identify that both processes could explain the result. Her next teaching move also stood in contrast to the implementation of this activity in the prior semester. Whereas she had asked students in semester one to identify which of two plots represented a particular process, Irene challenged students in semester two to determine how to experimentally differentiate between the two processes before she showed hypothetical data. Irene then prompted students to “have a discussion.” Reflecting on this open-ended question, Irene told us that: It's a good question because it's not the kind of question I've asked before very much. I've realized that this semester I am doing more of this experimental design… There wasn't just a single right answer… I felt they were understanding more than I have in the past. Were all the answers perfect? No, but I think getting them to think [about] how they would make that decision, it's a more creative process. They're having to engage in a higher-level learning ability, trying to critically think, analyze, and then create. I suppose that's what I'm trying to do is give those small opportunities to really get creative. Because this is a large class, I can't give them projects that I can grade. I can't do short answer long answer kind of questions. So, it's like this is the opportunity for them to be a little more creative. (Irene) Irene viewed the open-ended questions that she added as opportunities for generative work, which she described as “creative” in nature. Irene considered these opportunities valuable even if not all students ultimately arrived at a correct answer and felt they were feasible even in her large class. Irene summed up her thinking on her new approach, Learning is not just memorization. There are many other forms of learning. Being able to apply your knowledge in various ways is really important. [Students] have a deeper understanding of the subject if they can apply and be creative with that knowledge. (Irene) Irene's teaching practices changed in the second semester to include additional opportunities for students to discuss their reasoning with their peers after an initial poll. She also changed a closed-ended question to an open-ended question involving experimental design. These adjustments align with her developing understanding that generative work and the process of engaging students in explaining their reasoning fosters deeper learning. We also learned that at least two different teaching professional development experiences influenced Irene's thinking and teaching practice. Irene's case demonstrates how pedagogical knowledge of how people learn can influence nuanced aspects of active-learning instruction and can be supported by development opportunities. Case 2: Dana develops and uses pedagogical knowledge about refraining from explaining. In her upper-division developmental biology course, Dana's target lesson on the mammalian sex determination pathway consisted entirely of students working on a challenging set of problems. Dana's course included several lessons like this, which she called “pathway days,” and she considered them an important counterpart to lessons that prioritized lecture with some in-class questions. She designed pathway days to “really delve into” something from prior lessons so students “can work through it step-by-step and understand something that is yes, a little bit complicated or maybe something that is an offshoot but enforces other concepts that we've learned through the semester.” In the target lesson, Dana gave students a schematic of the steroid hormone biosynthesis pathway showing how, step-by-step, various enzymes facilitate the production of hormones and other molecules important for sex determination. During class, students worked through eight multiple choice questions. Neither the questions nor the pathway schematic changed at any point during four semesters of data collection, but as described in this case, Dana changed the way she facilitated students’ work on these questions over time. The eight questions in the target lesson varied from engaging students in generative work (e.g., identifying probable outcomes of perturbations in the pathway) to recalling processes and information (e.g., remembering which hormone is produced by the testes). In all four semesters, Dana interacted with students for more than 90% of the target lesson as they worked in small, informal groups. However, despite interacting one-on-one with small groups for most of the class, students explained their reasoning to Dana in less than 10% of two-minute lesson segments in semesters one, two, and three of the study. Though she interacted extensively with students in these semesters, Dana spent most of this time explaining concepts to students rather than inviting their explanations and reasoning. When she talked to small groups in these initial three semesters, Dana primarily asked students to recall information, using questions like, “What hormonesl do the testes make?” and then explained her own logic for how to work through the pathway to answer questions (refer to examples of these interactions from each semester in Supplementary Table S2). This approach to small group interactions was at odds with her pedagogical knowledge of how people learn. As early as the first semester, Dana made it clear that she viewed opportunities for students to explain their reasoning as important moments for learning (i.e., To explain is to learn ) when she said, Forcing them to talk it out, I think, is useful because, even talking to yourself, going with the idea of like, ‘You don't know something until you have to teach it’… I think just forcing them to put their logic in words. I think you can kind of tell your story in your head, but when you say things out loud—like when we say, ‘Oh my gosh, I just heard what I said.’ It's that type of processing that you're not only saying it, but you hear it at the same time. I think just talking it through out loud can be very helpful, even if you're just talking to yourself. (Dana) However, as Dana interacted with small groups, her actions ushered students to the right answer rather than eliciting their reasoning, thereby missing opportunities to let students learn by explaining their reasoning to her. In the second, third, and fourth semesters, Dana displayed a new pedagogical idea: Refraining from explaining keeps the onus on students to engage in generative work. Dana described her strategy when working with small groups this way: “as much as possible, I try to ask questions.” She also reflected on the tension she experienced between explaining to students and creating space for students to explain their thinking to her, I'll admit, that's something I'm trying to be better about… I think sometimes it's definitely easier to just, ‘Let me just explain this to you, instead of really making you work through it.’ (Dana) Dana pointed out that her role as an information resource in the classroom is a double-edged sword during pathway days because: … for some students, I think it would be a beneficial thing to be able to grab me and… clear something up a little more quickly. For other students, I think I might serve more as a crutch. (Dana) Dana was aware that students sometimes needed explanations but also recognized that students might use the instructor to bypass cognitive work important for their learning during class. Dana discussed how resisting the temptation to explain is challenging when students’ reasoning is incorrect but that showing this kind of restraint was crucial for her student's learning. Commenting on a longer than usual pause after she asked a challenging question to a group of students, Dana explained, One of the challenges with this is trying not to talk, trying to wait it out. Because again, I can say it all I want, but I'm not the one who needs that practice verbalizing it. I already know what I'm talking about. (Dana) Dana implied that sometimes remaining silent is the best instructional move a teacher can make. An important aspect of Dana's thinking that seemed to make her comfortable with waiting for students to respond was her understanding that, as a content expert, she did not need to practice explaining her reasoning. Instead, she thought these opportunities for learning should be left for students. Alongside this new idea, Dana described changes she planned to make to her teaching practices. She aimed: … to prompt [students] to explain your thought process on this. And I think I do a lot more of, you know, ‘Well, why is that? Like, why do you think that, not that it's right or wrong. You're right. But just explain that to me so I can try and get at what the process is, not just the endpoint.’(Dana) Dana also continued to experience tensions between her instructional goals and perceived limitations, I will admit it. I don't always do that. Sometimes, I launch into explanations because other people are waiting or things like that. It's just the time logistics. But, in an ideal world where I have the time, it's ‘Okay, talk me through your answers. Why do you think this one's correct? Why do you not think that one is correct?’ (Dana) Prior to Dana's fourth and final semester, she used her planned lead-in prompt, “Talk me through your logic” a handful of times, which successfully elicited students’ reasoning. However, after this lead-in prompt, Dana nearly always proceeded to do most of the explaining for the remainder of the interaction, asking students only basic recall questions (refer to Supplementary Table S2). Systematic video analysis of her instructional practices indicated no change in how often she heard students’ explanations in the second or third semester. In the fourth and final semester, Dana continued to demonstrate the pedagogical idea Refraining from explaining , and her teaching practices changed to align with this knowledge. Dana heard students explaining their reasoning in roughly 50% of 2-min lesson segments, which was the highest percentage of class time spent hearing reasoning across all participants and all time points in the study, and a five-fold increase from Dana's prior target lessons (Supplementary Figure S1). Looking closely at her interactions with small groups, Dana often refrained from explaining and instead facilitated generative work by asking questions that invited reasoning. Dana frequently took time to ask students to clarify or restate what they had said. Dana prompted students to make predictions and when students provided their prediction without accompanying reasoning, she followed up to ask, “Why?” These nuanced behaviors during small-group interactions marked a departure from semesters one, two, and three, where Dana did the bulk of the cognitive work by explaining the content and asking a series of pointed recall questions to walk students through the logic needed to reach an answer. In semester four, Dana kept the onus on students to do the generative cognitive work. Even though Dana developed ideas about facilitating generative work in her second semester of the study (fourth semester teaching college), we did not see Dana implementing this knowledge as part of her teaching practices until two semesters later. We do not have causal evidence to explain why this development in her practices occurred later, but the sequence is consistent with a scenario where this knowledge development informed changes to her practice. We may have observed a delay in timing between the pedagogical idea and the practice for a number of reasons, including the fact that the lesson was held virtually in semesters two and three (Supplementary Figure S1). An alternative explanation could be that it simply took time to develop the skilled practice of consistently eliciting student reasoning during class. DISCUSSION The central findings of this research are that (1) college faculty who use active learning rely on a range of generative and nongenerative pedagogical ideas about how people learn in their instructional decisions, and (2) they do not necessarily develop more generative pedagogical knowledge through teaching experience alone. The cases also demonstrate that pedagogical knowledge has the potential to influence teaching practices. This finding is important because nuanced instructional practices impact the learning opportunities created for students (e.g., Knight et al. , 2013 ; Knight et al. , 2015 ). Studies examining the details of how instructors implement evidence-based strategies find that instructors modify critical components—those that make the strategy effective for student learning—in ways that undermine the impact of the strategy (e.g., Dancy et al. , 2016 ). For example, 80% of physics instructors who use peer instruction do not consistently ask students to formulate an answer individually before talking in small groups, thereby taking away a critical opportunity for each student to engage in generative work ( Dancy et al. , 2016 ). Variation in pedagogical knowledge is a promising explanation for why some instructors make adaptations that render strategies less effective (e.g., Smith, 2015 ). Below, we link pedagogical ideas used by our participants to relevant research. Supporting the development and use of this pedagogical knowledge could equip instructors to adapt evidence-based strategies for their context in ways that maintain the productive learning experiences they hope to create for students. We discuss how instructors could be supported to develop this knowledge, propose implications of this research for college STEM instructors, and discuss limitations. Participants developed and relied upon promising generative pedagogical ideas Collectively, participants developed and relied on a productive suite of pedagogical ideas about how people learn that have the potential to inform the design and facilitation of generative learning opportunities for students. This section discusses each idea in light of relevant related research. The pedagogical idea Application and practice aligns with contemporary recommendations that scientific literacy involves developing deep conceptual understanding and engaging in the practices of sciences (e.g., American Association for the Advancement of Science [AAAS], 2011 ; National Research Council [NRC], 2012 ). Participants wanted students to be able to use the concepts and skills they learned in the course and therefore created opportunities for students to practice this during class. This represents a marked departure from the traditional view of college science classrooms as transmitting knowledge via lecture. In a study of math and science faculty's ideas about student learning, two of the most common ideas were that learning requires practice and perseverance and hands-on application ( Hora, 2014 ), suggesting that our finding has relevance beyond life sciences instructors. Participants described different goals for engaging students in the work of application and practice in class, including giving students the opportunity to use what they learned to “solve problems in novel scenarios” (i.e., predicting effects of a mutation in a biochemical pathway) and giving students practice with applying accurate use of terminology (i.e., identifying which element of a neuroscience experiment is the conditioned stimulus in a previously unseen experiment). The latter represents a way of thinking about applying knowledge that emphasizes terminology and therefore may not support the design of tasks that engage students in generative work. Notably, it was uncommon for participants’ pedagogical knowledge to focus on engaging students in the practices of science, though there have been national calls for this shift for over a decade ( AAAS, 2011 ; NRC, 2012 ) and numerous research-based resources exist to support instructors in incorporating science practices into their instruction (e.g., Laverty et al. , 2016 ; Bain et al. , 2020 ; Clemmons et al. , 2020 ). Based on our findings and research about the value of generative work for student learning, we encourage college biology instructors—at all levels—to replace application-style questions focused on terminology with questions that challenge students to apply conceptual knowledge to novel scenarios and to engage in science practices. Additionally, future research should consider how science instructors’ pedagogical knowledge could be leveraged and advanced to support instruction that prioritizes scientific practices. Instructors relied on the pedagogical idea To explain is to learn in making decisions about core active-learning strategies in their courses, including group work and individual writing. This idea recognizes that externalizing one's ideas, whether verbally, in writing, or graphically, requires and contributes to deep understanding. In previous work, 45% of life sciences faculty expressed the idea that students learn by verbally articulating their understanding, whereas only 29% of the whole sample of math and science faculty shared this idea ( Hora, 2014 ), potentially indicating that explaining as a tool for learning is particularly salient in the life sciences. Participants who expressed this idea often opted to use class time for small group work to create opportunities for students to explain their reasoning to peers. This is promising because student-to-student discussions about challenging questions positively impacts student learning, but instructors often skip this part of active-learning strategies ( Vickrey et al. , 2015 ). For instance, about 40% of physics faculty using peer instruction reported that they did not consistently have students discuss ideas in groups ( Dancy et al. , 2016 ). Drawing on the pedagogical idea To explain is to learn could help instructors feel comfortable allotting several minutes of class for students to discuss a question, as they would know that explaining one's thinking is essential for deep learning. To explain is to learn could also inform the guidance that instructors provide to students about how to discuss a question with peers. When instructors verbally direct students to focus on the reasons for their answers in small-group discussion, their discussions include more reasoning and connections to evidence than when verbal directions do not mention reasoning ( Knight et al. , 2013 ; Knight et al. , 2015 ). Thus, the way an instructor verbally frames their instructions for small group work has the potential to improve learning by prompting students to engage in generative work, and the pedagogical idea To explain is to learn could support verbal directions focused on explaining reasoning. Participants shared the pedagogical idea Generative work is a mirror as a mechanism by which opportunities for generative work supported student learning. This pedagogical idea relates to the concept of metacognition. Metacognition is a student's awareness and control of thinking for learning, and it can positively influence learning ( Cross and Paris, 1988 ; Wang et al. , 1990 ; Schraw and Moshman, 1995 ; Ohtani and Hisasaka, 2018 ). It includes metacognitive knowledge, what we know about our thinking, and metacognitive regulation, how we control our thinking for the purposes of learning ( Dye and Stanton, 2017 ). Some participants viewed generative learning opportunities as a way for students to recognize what they know and what they do not know, which is an opportunity to build metacognitive knowledge. By engaging in the process of self-explaining, students generate inferences that help them deepen their understanding (e.g., Chi et al. , 1994 ). Participants sometimes associated the metacognitive benefits of generative work (i.e., students’ recognition of what they know and do not know) with the moment in class where a correct answer was revealed, whereas other participants expected students to build metacognitive knowledge through the process of articulating their reasoning to others. This distinction may be important because an instructor with the first idea might prioritize the revelation and discussion of answers, whereas an instructor with the second idea might prioritize class time for students to explain their reasoning to each other and the class. The nuance of this pedagogical idea and its implications for fine-scale instructional decision-making warrant further investigation in future work. Importantly, students need metacognitive regulation skills to act on their metacognitive knowledge to support their learning. Instructors in this study did not share pedagogical knowledge about supporting students’ metacognitive regulation, but it is possible that this knowledge simply was not salient in interviews. Previous work suggests that undergraduates may need help from instructors to modify their learning strategies to address gaps in understanding (i.e., metacognitive knowledge) that they develop through generative work in class ( Stanton et al. , 2019 ; Halmo et al. , 2022 ; Halmo et al. , 2024 ). Stanton et al. , (2021) provide an evidence-based teaching guide about student metacognition for instructors who want to support students’ metacognitive skills alongside creating opportunities for generative cognitive engagement. The pedagogical idea Refrain from explaining was the rarest generative idea among participants, and yet it may serve a crucial function in helping instructors maintain high-quality learning opportunities for students. Recent research in K12 science education distinguishes between the design of a challenging cognitive task and how a task is launched and facilitated by the instructor ( Tekkumru-Kisa et al. , 2019 ; Tekkumru-Kisa et al. , 2020 ; Tekkumru-Kisa et al. , 2022 ). Each of these stages impact the learning opportunities created for students. A task with the potential for prompting generative work may not actually do so if the launch of the task focuses on following procedures to reach an answer (i.e., doing school) rather than reasoning about a phenomenon to figure out how and why (i.e., doing science; Tekkumru-Kisa et al. , 2020 ). Even if an in-class activity has been designed and launched to create opportunities for generative work, the onus for students to do this work can be maintained or compromised depending on the nuances of an instructor's facilitation of small-group and whole-class discussions. One of the main ways in which instructors undermine opportunities for generative work is by taking over the cognitive work themselves as they share their own explanations with students instead of letting students learn by explaining their thinking ( Stein et al. , 1996 ; Kang et al. , 2016 ). Dana's case illustrates that intentionally withholding explanations can be hard to do, even for those who consistently demonstrate the pedagogical idea Refrain from explaining , but that it is an area of teaching expertise that can be developed. Future work, including research and educational development innovations, can build on the literature regarding wait time during teacher-student discourse by examining what additional knowledge, skills, and support college instructors need to withhold their own explanations and instead prompt students to elaborate on their thinking ( Ingram and Elliot, 2016 ). This work could extend a line of K12 research into higher education contexts where it remains underexplored. We propose that three of the generative pedagogical ideas identified in this study are particularly complimentary and impactful to effective active-learning instruction: To explain is to learn , Generative work is a mirror, and Refraining from explaining. These three ideas revolve around students learning from explaining their reasoning. We propose that together these ideas are fruitful guiding principles for designing, launching, and facilitating generative work ( Stein and Smith, 1998 ; Tekkumru-Kisa et al. , 2019 ; Tekkumru-Kisa et al. , 2020 ; Tekkumru-Kisa et al. , 2022 ). The pedagogical idea To explain is to learn can guide the design and launch of in-class tasks that are challenging enough to require an articulation of reasoning. Similarly, the idea Generative work is a mirror can inform the design and launch of in-class tasks that specifically prompt students to monitor what they do and do not know as they engage with task. Lastly, the idea Refraining from explaining can help instructors withhold their own explanations as they facilitate small-group and class discussions until students have had the opportunity to articulate their reasoning. Despite the complementarity of these pedagogical ideas, it was rare for participants to use all three of these ideas concurrently. Only two out of eleven participants (18%) demonstrated evidence of all three ideas at a single time point. More commonly (7/11; 64%), participants demonstrated two of these ideas at the same time point. Future work can determine the impact of capitalizing on all three of these ideas, relative to one or two, on instructional practices and student outcomes. Nongenerative pedagogical ideas remained salient for participants Participants consistently used a few pedagogical ideas aligned with lower levels of cognitive engagement, even after they developed generative ideas. In particular, the idea Breaking up lecture was expressed by every participant in every semester and the idea Peer-to-peer listening was nearly as common, with both occurring alongside generative ideas ( Figure 2B ). The ubiquity of these ideas across participants suggests that these ideas are intuitive or commonly developed through previous experiences, perhaps as learners or as college instructors. This study could not uncover the origin of these ideas, but another study of 53 STEM faculty indicated that about twice as many faculty drew on their teaching experiences compared with their experiences as students to inform their own teaching ideas ( Oleson and Hora, 2014 ). The persistence of these ideas over time, even among instructors who expanded their knowledge, suggests that they are reinforced or remain functionally useful (e.g., Gouvea, 2023 ). Future work that contrasts instructors with developed pedagogical knowledge of how people learn and instructors whose pedagogical knowledge is still developing could help determine whether and how the active and passive pedagogical ideas identified in this study remain salient for instructors as they gain experience, what other knowledge informs these ideas, and whether they provide any unique benefits or liabilities when it comes to fostering student learning. Development of pedagogical knowledge of how people learn An important question emerges from this study: what experiences foster the development of pedagogical ideas about how people learn? Though our longitudinal design allowed us to document the trajectory of participants’ pedagogical knowledge, it does not necessarily illuminate the origin of new ideas. Teaching professional development (PD) experiences seem like the most obvious opportunities for developing new teaching knowledge, and participants engaged in teaching PD extensively. Irene's case provides a specific example of how teaching PD led to a new pedagogical idea and related practice. She credited a workshop with expanding her sense of what kinds of cognitive work she can facilitate in a large lecture hall, prompting her to move beyond multiple choice questions toward open-ended generative tasks that promote metacognition. However, other participants who engaged in extensive teaching professional development did not develop new pedagogical knowledge of how people learn and did not substantially change their active-learning instruction. This seems to suggest that teaching professional development experiences may not meaningfully shifting faculty knowledge about how people learn. In the current system of higher education, participation in professional development is almost always voluntary, and not necessarily encouraged (e.g., Lane et al. , 2019 ). Most opportunities for teaching development focus on equipping instructors with specific teaching practices (e.g., active learning, scientific teaching, alternative grading) and do not address the knowledge of how people learn needed to guide instructional decision-making as instructors plan and implement the focal practice (e.g., Henderson et al. , 2011 ). Our work suggests that teaching PD programs may need to focus on pairing the instructional strategies they intend for participants to use with opportunities for participants to engage in generative work to understand the mechanisms of learning underlying the strategies and how that relates to adaptations one might make to the strategy. In doing so, teaching PD should also consider the likelihood that pedagogical reasoning is context-dependent, meaning that participants may develop principles of pedagogical knowledge they apply when planning instruction and yet struggle to apply those principles as they facilitate an in-class activity. For instance, as Dana's case demonstrated, instructors can apply the idea of Refraining from explaining in the planning phase while struggling to uphold that principle during interactions with students. Given the limited opportunities for undergraduate instructors to engage in teaching PD, it is notable that certain teaching experiences provide opportunities to build pedagogical knowledge. However, we caution against the assumption that experience alone is sufficient for expertise development. Not all participants, even highly motivated instructors who pursued extensive teaching professional development, built pedagogical knowledge about generative work during this study. This point is critical to emphasize because systems of higher education are built on the rarely scrutinized assumption that instructors will learn how to teach effectively based solely on their experience as faculty. The observer will find this assumption laid bare in the lack of formal pedagogical training required of early-career undergraduate STEM instructors (e.g., Schussler et al. , 2015 ), the lack of formalized teaching mentorship programs available to faculty (e.g., Brickman et al. , 2016 ), and the lack of formal opportunities for instructors to receive regular and constructive feedback on their instruction (e.g., Brickman et al. , 2016 ). Additionally, co-teaching, in which an instructor plans, teaches, and reflects on teaching in close collaboration with a colleague, may be a mechanism for helping faculty develop expert teaching knowledge and practice (e.g., Henderson et al. , 2009 ; Bailey and Nagamine, 2012 ; Scherer et al. , 2020 ; Haag et al. , 2023 ). We observed an example of this in our dataset. Beth demonstrated an understanding of generative work as a mirror during her first semester as a college teacher and incorporated a reflective writing exercise into her class; she credited her co-teacher for both this practice and her related knowledge. However, “co-teaching” in science often involves multiple instructors teaching segments of a class with little coordination. This arrangement is unlikely to provide opportunities for instructors to hear one another's rationale for teaching decisions and may not be any more effective at fostering knowledge development than teaching alone ( Haag et al. , 2023 ). Our data also suggest a third catalyst for new pedagogical knowledge: instructor self-reflection. Self-reflection can contribute to teaching professional expertise development by helping instructors surface and interrogate the tacit knowledge that informs their instructional decision-making. This kind of reflection creates opportunities for instructors to recognize differences between what they say they do in their teaching and what they apply in the classroom (Schon, 1987; Marcos et al. , 2009 ). Teaching reflection can be a process of continual inquiry into one's practice that relies on and contributes to one's pedagogical knowledge ( McAlpine et al. , 1999 ). It can involve making plans to change some aspect of teaching practice; enacting the change; and monitoring information from learners, observations, and other cues to evaluate whether the change met the goals (e.g., McAlpine et al. , 1999 ; Marcos et al. , 2009 ). Here we propose that instructors might develop useful pedagogical knowledge by reflecting on the continued learning of course content that they experience in the process of teaching. This might be relatively simple, such as by asking faculty to reflect on when they felt that they had the strongest grasp of the content they teach, what contributed to their deep understanding, and how might they be able to create similar learning opportunities for their students in their own classroom. Though teacher reflection has long been a considered in teacher development in K12 educational contexts, there is less work on how college instructors develop the skills for engaging in this kind of reflection (Kane et al. , 2004; Marcos et al. , 2009 ). Instructors likely need ongoing support and feedback as they work to apply new insights in practice, determine what worked and what did not, and make further plans to refine their approach. We observed numerous examples of instructors justifying the pedagogical idea To explain is to learn as they reflected on their own experiences of learning through the act of teaching. For example, Kim explained, “I think that's one of the reasons why teaching helps you understand material so much better is because you have to think about why something makes sense to you and you have to think about what about it doesn't make sense to you in order to try to explain it clearly to somebody else. So, I think that process is how you really learn.” Like several other participants, Kim had come to recognize that the process of explaining content to her students deepened her own learning of the content as an instructor. Participants sometimes connected their own learning that occurs when they teach to the learning that students experience when they explain their thinking to their peers. For instance, June made this connection, drawing on both her own experiences as a student and her observations of students in her class, “I think when I was a student, I learned the most by not only being able to figure it out myself, but also being able to teach it to other people. So, for the students who do understand it and are getting it, no problem. They can then teach it to their neighbor and reinforce that, or as they teach it, they realize, ‘Oh, I have a misconception there’, or ‘I have a lack of understanding of this section.’” Implications for instructors In this study, participants displayed a range of pedagogical knowledge of how people learn that aligned with different levels of the ICAP framework. Each of these ideas has the potential to be productive starting places for college instructors as they seek to further develop their pedagogical knowledge. With that in mind, and positioning faculty as learners, we have outlined a set of initial pedagogical ideas, based on our data, that we suspect may be present in the broader undergraduate STEM instructor population ( Table 4 ). We link these ideas to reflection questions that might be helpful for STEM instructors as they seekto develop their knowledge of how people learn. These questions could also be a resource for educational developers as they plan programming to support the adoption and effective use of active-learning strategies ( Table 4 ). We also direct readers to Supplementary Table S3, which elaborates on Table 4 by providing quotes from participants that exemplify more developed pedagogical knowledge corresponding to each initial idea. We refrain from including these quotes in the main paper to preserve the opportunity for instructors to engage in generative work as they respond to reflection questions with their own ideas. These questions could be answered in writing or in a collaborative dialogue with a teaching colleague. TABLE 4: Initial ideas about how people learn that were common among participants and reflection questions to support instructors in building upon these ideas. Supplementary Table S3 includes an expanded table with examples of pedagogical ideas from participants that build upon initial ideas. Initial pedagogical idea Reflection questions Active learning is valuable because it keeps students attentive and engaged throughout a lesson. “[Active learning] makes them stop and actually do something. So, that's going to just be a change in their attention. So, they wake up and they're not just like passively sitting there the whole time, they're actually being called out to do something.” – Fiona What, besides keeping students’ attention, do I aim to achieve with active-learning strategies? And which of these goals are most important to me? What aspects of my active-learning instruction rely on, or are consistent with, the assumption that students primarily learn from listening to lectures? Group work is valuable because students have the chance to listen to how their peers explain concepts. “I think they hear a different voice. I think it's that simple. I think hearing it from somebody else who's learning it the first time they might have a different way of explaining it or, or like thinking about it themselves that might resonate with one of them.” – George What, besides hearing each other's ideas, do I hope students gain from in-class group work? What would I consider an “ideal” small-group discussion for fostering learning? What intentional directions could I provide to students prior to small group work to support them in creating an “ideal” discussion? Active learning works because students have to write down or explain concepts in their own words. “… then the pair and share part, you know, I'm asking them to discuss with a partner, so that if they are totally confused, they can get help. But also, if they're not sure, but they have a good idea in sharing they'll boost their confidence, I guess. And so, I think that that helps them practice their answer. Then, when I ask them to speak aloud in the class, they're better prepared to do so because they've run it by one person before. So, they might have more confidence that they have a good idea.” – Beth What is it about writing or explaining one's thinking that supports student learning? What could I do or say to help students learn as much as possible from an opportunity to write or explain their own thinking? When I interact with students while they work, I try to prevent them from getting stuck by asking questions to lead them through a logical progression. “If I had all the time in the world, I would baby step them and say like, okay, what is that part of the image representing? And okay, as a result, what's it doing here? … Like walking them step-by-step of like, ‘Okay, let's focus over here first. Let's make sure you understand that before we move on to how that impacts whether it's transcribed or not is what I would've loved to do.” – June If time were not a limitation, how would you want to balance between asking questions that point students toward an answer and letting them struggle to construct an answer on their own? For the activities in an upcoming lesson, which questions from students might I answer directly, which will I refrain from explaining, and why? Open in a new tab Limitations We urge readers to carefully consider the study's limitations in interpreting findings and considering implications. First, our sample of participants differs from the typical undergraduate instructor population in several important ways, limiting the generalizability of our findings to all biology faculty. Notably, many of our participants had a high level of engagement in teaching professional development, with the majority having accumulated over 40 h of such training and a subset having experience in education research. This level of commitment suggests that these instructors were particularly invested in improving their teaching and likely had more opportunities to develop pedagogical knowledge than the average faculty member. Another potential limitation is the impact of the COVID-19 pandemic on our longitudinal data collection. The pandemic forced many instructors to adapt their target lessons to fit new instructional modalities, which influenced their teaching strategies and potentially the pedagogical knowledge they shared as they discussed the target lessons. We took steps to mitigate the impact on the study by prompting participants during interviews to avoid discussions of COVID-related instructional changes. Additionally, we examined how these shifts in modality related to the pedagogical knowledge demonstrated by participants and did not detect patterns that would invalidate our findings (refer to Supplementary Figure S2). Interviews in this study provided opportunities for teaching self-reflection that participants may not have otherwise experienced. In fact, participants sometimes commented that they had not had previous opportunities to reflect on their teaching as deeply as they did during interviews. This could have led to knowledge and practice development that might not have otherwise occurred. This limitation was difficult to avoid given our in-depth interviews and longitudinal design. It requires considerable intentionality to make time to reflect on one's teaching and it is not typically an expectation of faculty positions. Yet, reflecting on the purpose of a particular strategy can help an instructor recognize implicit goals and underlying assumptions that may be in play and point toward priorities for implementing the strategy to ensure it meets the intended goals. If the opportunity to engage in the study supported more knowledge development than otherwise would have occurred, it further emphasizes the finding that teaching experience alone is insufficient to promote pedagogical knowledge of how people learn. The ICAP framework provided useful theoretical grounding for this study and could also be a resource for instructors looking to refine their active-learning practice ( Chi et al. , 2018 ). Practitioners interested in leveraging the ICAP framework in their teaching should be made aware of an important assumption embedded in the framework. The overt student behaviors associated with each level of cognitive engagement are assumed to be indirect indicators of covert learning processes. As Chi (2021) points out, the mapping of covert learning processes to overt student behaviors is imperfect. This is because students can covertly engage in generative cognitive work while outwardly exhibiting active and passive behaviors. As such, relying on students’ overt behaviors as the only evidence of learning will overlook covert learning that occurs in the classroom. Rather than assuming students’ overt behaviors represent their thinking processes, instructors should also elicit student thinking during class to get a real-time, formative assessment of student learning (e.g., Gehrtz et al. , 2022 ; Thurn et al. , 2023 ; Waugh et al. , 2025 ). Readers should be aware that the construct of pedagogical knowledge has some overlap with conceptualizations of teacher beliefs and orientations. Researchers seem to agree that both knowledge and beliefs can influence instructional practices (e.g., Southerland et al. , 2001 ; Luft and Roehrig, 2007 ), yet it is often unclear what would distinguish knowledge from beliefs. Some researchers view beliefs as more emotionally laden and subjective than knowledge, more likely to be derived from life experience rather than formal training, or lacking an evidentiary basis (e.g., Luft and Roehrig, 2007 ). However, these features might also be present in nascent ideas of someone without formal training in a domain, so we did not find this distinction useful for our work. Orientations are general ways of viewing science teaching and are also difficult to distinguish from conceptualizations of beliefs (e.g., Magnusson et al. , 1999 ; Friedrichsen et al. , 2011 ). Contemporary theory positions beliefs and orientations as acting as amplifiers and filters between knowledge and practice and recognizes pedagogical knowledge as a distinct teacher professional knowledge base ( Gess-Newsome, 2015 ; Carlson et al. , 2019 ). As such, we interpret all pedagogical ideas uttered by participants in this study as professional knowledge. We see neither a compelling reason nor sufficient evidence to argue that knowledge is clearly distinct from beliefs or orientations. Other researchers could apply other interpretive frameworks that position certain utterances as beliefs or orientations to our data. These alternative interpretations are not likely to undermine the novelty or contributions of this work, as our findings and their implications would likely remain similar whether the qualitative evidence underpinning our claims were interpreted as knowledge, beliefs, or orientations. We take the stance that teachers are life-long learners in the classroom and find that referring to their pedagogical ideas as “knowledge” recognizes the intellectual work of teaching and implies that their expertise can be expanded over time. Finally, our study focused on a single target lesson to characterize instructors’ pedagogical knowledge and practices. If we had observed a different lesson, our findings could be different. For the purposes of our longitudinal design, this focus was key to observing changes to lesson design, framing, and implementation over time, but limits the generalizability of findings beyond the target lesson. Participants’ knowledge of how students learn the specific topics included in their target lesson and how to teach them, their pedagogical content knowledge, likely shaped how they developed and applied their general pedagogical knowledge of how people learn across different lessons. Although this study focused on pedagogical knowledge of how people learn, we recognize that a teacher's topic-specific pedagogical content knowledge is also essential to effective instruction and interacts meaningfully with the ideas presented here. In a separate study of the same data set, we examine how the same sample of instructor participants developed pedagogical content knowledge and how this knowledge development influenced their teaching practices ( Waugh et al. , 2025 ). CONCLUSION Research on evidence-based teaching practices often emphasizes the strategies themselves, which can overlook the foundational knowledge instructors need to use these practices effectively. This study identified seven distinct pedagogical ideas that instructors relied on in their design and implementation of active learning, some of which are better aligned with theory and research about how people learn. Not all participants exhibited clear patterns of developing pedagogical knowledge over time, raising important questions about the common assumption that college instructors can effectively learn on-the-job without external support. We encourage teaching professional developers and researchers to consider and prioritize teaching knowledge, ensuring that publications of evidence-based strategies include the necessary resources for instructors to develop critical teaching knowledge (e.g., Stains and Vickrey, 2017 ). Ultimately, the success of evidence-based practices in improving student outcomes relies on their intentional implementation. Thus, advancing STEM education will require supporting faculty in developing teaching knowledge and rewarding those who pursue this type of professional growth. Supporting information cbe-25-ar11-s001.pdf (520.5KB, pdf) ACKNOWLEDGMENTS We extend our gratitude to our research participants for their valuable time and energy. They stuck with the study over repeated semesters and through a global pandemic, and we feel honored to have had the chance to learn alongside them. We are also grateful to the University of Georgia's Biology Education Research Group and the Andrews lab for their feedback and support. Dr. Sandhya Krishnan trained an initial undergraduate team on our observation protocol. Aman Patel led his fellow undergrads in implementing the protocol, including Esha Mohnalkar, Michella Obialor, Katia Uriostegui-Santos, and Bryn Robinson. We thank two anonymous reviewers and the editor for helpful feedback that improved the quality of this manuscript. Support for this work provided by the National Science Foundation's Improving Undergraduate STEM (IUSE) and ECR program under Award no. 1845886. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. Footnotes 1 Auerbach and Andrews (2018) and subsequent papers refer to “knowledge of how people learn” as “knowledge of opportunities for legitimate generative work”. Here, we adopted the simpler and more general name “knowledge of how people learn” to encompass a broader range of ideas about learning, but the two constructs overlap. REFERENCES Alonzo, A. C., Kim, J. (2016). Declarative and dynamic pedagogical content knowledge as elicited through two video-based interview methods. Journal of Research in Science Teaching 53(8), 1259–1286. [ Google Scholar ] Andrews, T. C., Auerbach, A. J. J., Grant, E. F. (2019). Exploring the relationship between teacher knowledge and active-learning implementation in large college biology courses. CBE—Life Sciences Education 18(4), ar48. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Andrews, T. M., Leonard, M. J., Colgrove, C. A., Kalinowski, S. T. (2011). Active learning not associated with student learning in a random sample of college biology courses. CBE—Life Sciences Education 10(4), 394–405. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Andrews, T. C., Speer, N. M., Shultz, G. V. (2022). Building bridges: A review and synthesis of research on teaching knowledge for undergraduate instruction in science, engineering, and mathematics. International journal of STEM education 9(1), 66. [ Google Scholar ] American Association for the Advancement of Science. (2011). Vision and change in undergraduate biology education: A call to action. Washington, DC. [ Google Scholar ] Anfara, Jr A. V., Brown, K. M., Mangione, T. L. (2002). Qualitative analysis on stage: Making the research process more public. Educational researcher 31(7), 28–38. [ Google Scholar ] Auerbach, A. J. J., Andrews, T. C. (2018). Pedagogical knowledge for active-learning instruction in large undergraduate biology courses: A large-scale qualitative investigation of instructor thinking. International journal of STEM education 5, 1–25. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Auerbach, A. J., Higgins, M., Brickman, P., Andrews, T. C. (2018). Teacher knowledge for active-learning instruction: Expert–novice comparison reveals differences. CBE—Life Sciences Education 17(1), ar12. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Bailey, J. M., Nagamine, K. (2012). Experiencing conceptual change about teaching: A case study from astronomy. American Journal of Physics 80(6), 542–551. [ Google Scholar ] Bain, K., Bender, L., Bergeron, P., Caballero, M. D., Carmel, J. H., Duffy, E. M., … Cooper, M. M. (2020). Characterizing college science instruction: The three-dimensional learning observation protocol. PLoS One 15(6), e0234640. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Ballen, C. J., Wieman, C., Salehi, S., Searle, J. B., Zamudio, K. R. (2017). Enhancing diversity in undergraduate science: Self-efficacy drives performance gains with active learning. CBE—Life Sciences Education 16(4), ar56. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Brickman, P., Gormally, C., Martella, A. M. (2016). Making the grade: Using instructional feedback and evaluation to inspire evidence-based teaching. CBE—Life Sciences Education 15(4), ar75. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Carlson, J., Daehler, K. R., Alonzo, A. C., Barendsen, E., Berry, A., Borowski, A., … Wilson, C. D. (2019). The refined consensus model of pedagogical content knowledge in science education. Repositioning Pedagogical Content Knowledge in Teachers’ Knowledge for Teaching Science, 77–94. [ Google Scholar ] Chan, K. K. H., Hume, A. (2019). Towards a consensus model: Literature review of how science teachers’ pedagogical content knowledge is investigated in empirical studies. Repositioning pedagogical content knowledge in teachers’ knowledge for teaching science, 3–76. [ Google Scholar ] Chan, K. K. H., Yung, B. H. W. (2018). Developing pedagogical content knowledge for teaching a new topic: More than teaching experience and subject matter knowledge. Research in Science Education 48, 233–265. [ Google Scholar ] Charmaz, K. (2006). Constructing grounded theory: A practical guide through qualitative analysis. Sage. [ Google Scholar ] Chase, A., Pakhira, D., Stains, M. (2013). Implementing process-oriented, guided-inquiry learning for the first time: Adaptations and short-term impacts on students’ attitude and performance. Journal of Chemical Education 90(4), 409–416. [ Google Scholar ] Chi, M. T. (2009). Active-constructive-interactive: A conceptual framework for differentiating learning activities. Topics in cognitive science 1(1), 73–105. [ DOI ] [ PubMed ] [ Google Scholar ] Chi, M. T. (2021). Translating a theory of active learning: An attempt to close the research-practice gap in education. Topics in Cognitive Science 13(3), 441–463. [ DOI ] [ PubMed ] [ Google Scholar ] Chi, M. T., Wylie, R. (2014). The ICAP framework: Linking cognitive engagement to active learning outcomes. Educational psychologist 49(4), 219–243. [ Google Scholar ] Chi, M. T., Adams, J., Bogusch, E. B., Bruchok, C., Kang, S., Lancaster, M., … Yaghmourian, D. L. (2018). Translating the ICAP theory of cognitive engagement into practice. Cognitive science 42(6), 1777–1832. [ DOI ] [ PubMed ] [ Google Scholar ] Chi, M. T., De Leeuw, N., Chiu, M. H., LaVancher, C. (1994). Eliciting self-explanations improves understanding. Cognitive science 18(3), 439–477. [ Google Scholar ] Chi, M. T., Kang, S., Yaghmourian, D. L. (2017). Why students learn more from dialogue-than monologue-videos: Analyses of peer interactions. Journal of the Learning Sciences 26(1), 10–50. [ Google Scholar ] Clemmons, A. W., Timbrook, J., Herron, J. C., Crowe, A. J. (2020). BioSkills Guide: Development and national validation of a tool for interpreting the Vision and Change core competencies. CBE—Life Sciences Education 19(4), ar53. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and psychological measurement, 20(1), 37–46. [ Google Scholar ] Collins JW, 3rd, O'Brien NP (Eds.). (2003). Greenwood Dictionary of Education. Westport, CT: Greenwood. [ Google Scholar ] Cross, D. R., Paris, S. G. (1988). Developmental and instructional analyses of children's metacognition and reading comprehension. Journal of educational psychology 80(2), 131. [ Google Scholar ] Dancy, M., Henderson, C., Turpen, C. (2016). How faculty learn about and implement research-based instructional strategies: The case of peer instruction. Physical Review Physics Education Research 12(1), 010110. [ Google Scholar ] Daubenmire, P. L., Bunce, D. M., Draus, C., Frazier, M., Gessell, A., van Opstal, M. T. (2015). During POGIL implementation the professor still makes a difference. Journal of college science teaching 44(5), 72–81. [ Google Scholar ] Driessen, E. P., Knight, J. K., Smith, M. K., Ballen, C. J. (2020). Demystifying the meaning of active learning in postsecondary biology education. CBE—Life Sciences Education 19(4), ar52. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Dye, K. M., Stanton, J. D. (2017). Metacognition in upper-division biology students: Awareness does not always lead to control. CBE—Life Sciences Education, 16(2), ar31. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Eddy, S. L., Hogan, K. A. (2014). Getting under the hood: How and for whom does increasing course structure work? CBE—Life Sciences Education 13(3), 453–468. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Freeman, S., Eddy, S. L., McDonough, M., Smith, M. K., Okoroafor, N., Jordt, H., Wenderoth, M. P. (2014). Active learning increases student performance in science, engineering, and mathematics. Proceedings of the National Academy of Sciences 111(23), 8410–8415. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Friedrichsen, P., Driel, J. H. V., Abell, S. K. (2011). Taking a closer look at science teaching orientations. Science Education 95(2), 358–376. [ Google Scholar ] Gehrtz, J., Brantner, M., Andrews, T. C. (2022). How are undergraduate STEM instructors leveraging student thinking? International Journal of STEM Education 9(1), 18. [ Google Scholar ] Gess-Newsome, J. (2015). A model of teacher professional knowledge and skill including PCK: Results of the thinking from the PCK Summit. In Re-examining pedagogical content knowledge in science education, ed. Berry A, Friedrichsen P, and Loughran J, New York: Routledge, 28–42. [ Google Scholar ] Gouvea, J. (2023). Processing misconceptions: Dynamic systems perspectives on thinking and learning. Frontiers in Education 8, 1215361 [ Google Scholar ] Grossman, P. L., Richert, A. E. (1988). Unacknowledged knowledge growth: A re-examination of the effects of teacher education. Teaching and teacher Education 4(1), 53–62. [ Google Scholar ] Guba, E. G. (1981). Criteria for assessing the trustworthiness of naturalistic inquiries. Ectj 29(2), 75–91. [ Google Scholar ] Haag, K., Pickett, S. B., Trujillo, G., Andrews, T. C. (2023). Co-Teaching in undergraduate STEM education: A lever for pedagogical change toward evidence-based teaching? CBE—Life Sciences Education 22(1), es1. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Haak, D. C., HilleRisLambers, J., Pitre, E., Freeman, S. (2011). Increased structure and active learning reduce the achievement gap in introductory biology. Science 332(6034), 1213–1216. [ DOI ] [ PubMed ] [ Google Scholar ] Halmo, S. M., Bremers, E. K., Fuller, S., Stanton, J. D. (2022). “Oh, that makes sense”: Social Metacognition in Small-Group Problem Solving. CBE—Life Sciences Education 21(3), ar58. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Halmo, S. M., Yamini, K. A., Stanton, J. D. (2024). Metacognition and self-efficacy in action: How first-year students monitor and use self-coaching to move past metacognitive discomfort during problem solving. CBE—Life Sciences Education 23(2), ar13. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Hausmann, R. G., Van De Sande, B., VanLehn, K. (2008). Shall we explain? Augmenting learning from intelligent tutoring systems and peer collaboration. In Intelligent Tutoring Systems: 9th International Conference, ITS 2008, Montreal, Canada, June 23-27, 2008 Proceedings 9, 636–645. Springer Berlin Heidelberg. [ Google Scholar ] Henderson, C., Beach, A., Famiano, M. (2009). Promoting instructional change via co-teaching. American Journal of Physics 77(3), 274–283. [ Google Scholar ] Henderson, C., Beach, A., Finkelstein, N. (2011). Facilitating change in undergraduate STEM instructional practices: An analytic review of the literature. Journal of research in science teaching 48(8), 952–984. [ Google Scholar ] Henderson, C., Dancy, M. H. (2009). Impact of physics education research on the teaching of introductory quantitative physics in the United States. Physical Review Special Topics—Physics Education Research 5(2), 020107. [ Google Scholar ] Hora, M. T. (2014). Exploring faculty beliefs about student learning and their role in instructional decision-making. The Review of Higher Education 38(1), 37–70. [ Google Scholar ] Ingram, J., Elliott, V. (2016). A critical analysis of the role of wait time in classroom interactions and the effects on student and teacher interactional behaviours. Cambridge Journal of Education 46(1), 37–53. [ Google Scholar ] Jackson, M. A., Moon, S., Doherty, J. H., Wenderoth, M. P. (2022). Which evidence-based teaching practices change over time? Results from a university-wide STEM faculty development program. International Journal of STEM Education 9(1), 22. [ Google Scholar ] Kang, H., Windschitl, M., Stroupe, D., Thompson, J. (2016). Designing, launching, and implementing high quality learning opportunities for students that advance scientific thinking. Journal of Research in Science Teaching 53(9), 1316–1340. [ Google Scholar ] Knight, J. K., Wise, S. B., Rentsch, J., Furtak, E. M. (2015). Cues matter: Learning assistants influence introductory biology student interactions during clicker-question discussions. CBE—Life Sciences Education 14(4), ar41. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Knight, J. K., Wise, S. B., Southard, K. M. (2013). Understanding clicker discussions: Student reasoning and the impact of instructional cues. CBE—Life Sciences Education 12(4), 645–654. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] König, J., Blömeke, S., Klein, P., Suhl, U., Busse, A., Kaiser, G. (2014). Is teachers' general pedagogical knowledge a premise for noticing and interpreting classroom situations? A video-based assessment approach. Teaching and teacher education 38, 76–88. [ Google Scholar ] Koretsky, M. D., Nolen, S. B., Galisky, J., Auby, H., Grundy, L. S. (2024). Progression from the mean: Cultivating instructors' unique trajectories of practice using educational technology. Journal of Engineering Education 113(2), 330–359. [ Google Scholar ] Kramarski, B., Dudai, V. (2009). Group-metacognitive support for online inquiry in mathematics with differential self-questioning. Journal of Educational Computing Research 40(4), 377–404. [ Google Scholar ] Laverty, J. T., Underwood, S. M., Matz, R. L., Posey, L. A., Carmel, J. H., Caballero, M. D., … Cooper, M. M. (2016). Characterizing college science assessments: The three-dimensional learning assessment protocol. PloS One 11(9), e0162333. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Lane, A. K., Hardison, C., Simon, A., Andrews, T. C. (2019). A model of the factors influencing teaching identity among life sciences doctoral students. Journal of Research in Science Teaching 56(2), 141–162. [ Google Scholar ] Landis, J. R., Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics 33, 159–174. [ PubMed ] [ Google Scholar ] Luft, J. A., Roehrig, G. H. (2007). Capturing science teachers’ epistemological beliefs: The development of the teacher beliefs interview. The Electronic Journal for Research in Science & Mathematics Education 11, 38–63. [ Google Scholar ] Magnusson, S., Krajcik, J., Borko, H. (1999). Nature, sources, and development of pedagogical content knowledge for science teaching. In Examining pedagogical content knowledge: The construct and its implications for science education, Dordrecht: Springer Netherlands, 95–132. [ Google Scholar ] Marcos, J. J. M., Miguel, E. S., Tillema, H. (2009). Teacher reflection on action: What is said (in research) and what is done (in teaching). Reflective Practice 10(2), 191–204. [ Google Scholar ] McAlpine, L., Weston, C., Beauchamp, C., Wiseman, C., Beauchamp, J. (1999). Building a metacognitive model of reflection. Higher Education 37(2), 105–131. [ Google Scholar ] Menekse, M., Chi, M. T. (2019). The role of collaborative interactions versus individual construction on students’ learning of engineering concepts. European Journal of Engineering Education 44(5), 702–725. [ Google Scholar ] Menekse, M., Stump, G. S., Krause, S., Chi, M. T. (2013). Differentiated overt learning activities for effective instruction in engineering classrooms. Journal of Engineering Education 102(3), 346–374. [ Google Scholar ] Morine-Dershimer, G., Kent, T. (1999). The complex nature and sources of teachers’ pedagogical knowledge. In Examining pedagogical content knowledge: The construct and its implications for science education, Dordrecht: Springer Netherlands, 21–50. [ Google Scholar ] National Research Council. (2012). A Framework for K-12 Science Education. National Academies Press. [ Google Scholar ] Nehm, R. H., Finch, S. J., Sbeglia, G. C. (2022). Is active learning enough? The contributions of misconception-focused instruction and active-learning dosage on student learning of evolution. BioScience 72(11), 1105–1117. [ Google Scholar ] Offerdahl, E. G., McConnell, M., & Boyer, J. (2018). Can I have your recipe? Using a fidelity of implementation (FOI) framework to identify the key ingredients of formative assessment for learning. CBE—Life Sciences Education 17(4), es16. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Ohtani, K., Hisasaka, T. (2018). Beyond intelligence: A meta-analytic review of the relationship among metacognition, intelligence, and academic performance. Metacognition and Learning 13, 179–212. [ Google Scholar ] Oleson, A., Hora, M. T. (2014). Teaching the way they were taught? Revisiting the sources of teaching knowledge and the role of prior experience in shaping faculty teaching practices. Higher education 68, 29–45. [ Google Scholar ] Peper, R. J., Mayer, R. E. (1986). Generative effects of note-taking during science lectures. Journal of Educational psychology 78(1), 34. [ Google Scholar ] Pollock, S. J., Finkelstein, N. D. (2008). Sustaining educational reforms in introductory physics. Physical Review Special Topics—Physics Education Research 4(1), 010110. [ Google Scholar ] Popova, M., Kraft, A., Harshman, J., Stains, M. (2021). Changes in teaching beliefs of early-career chemistry faculty: A longitudinal investigation. Chemistry Education Research and Practice 22(2), 431–442. [ Google Scholar ] Saldaña, J. (2003). Longitudinal Qualitative Research: Analyzing Change through Time. AltaMira Press. [ Google Scholar ] Scherer, H. H., O'Rourke, M., Seman-Varner, R., Ziegler, P. (2020). Coteaching in higher education: A case study of instructor learning. Journal of Effective Teaching in Higher Education 3(1), 15–29. [ Google Scholar ] Schraw, G., Moshman, D. (1995). Metacognitive theories. Educational psychology review 7, 351–371. [ Google Scholar ] Schussler, E. E., Read, Q., Marbach-Ad, G., Miller, K., & Ferzli, M. (2015). Preparing biology graduate teaching assistants for their roles as instructors: An assessment of institutional approaches. CBE—Life Sciences Education 14(3), ar31. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Shenton, A. K. (2004). Strategies for ensuring trustworthiness in qualitative research projects. Education for information 22(2), 63–75. [ Google Scholar ] Smith, G. A. (2015). Why college faculty need to know the research about learning. InSight: A Journal of Scholarly Teaching 10, 9–18. [ Google Scholar ] Smith, M. K., Jones, F. H., Gilbert, S. L., Wieman, C. E. (2013). The Classroom Observation Protocol for Undergraduate STEM (COPUS): A new instrument to characterize university STEM classroom practices. CBE—Life Sciences Education 12(4), 618–627. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Southerland, S. A., Sinatra, G. M., Matthews, M. R. (2001). Belief, knowledge, and science education. Educational Psychology Review 13, 325–351. [ Google Scholar ] Stains, M., Vickrey, T. (2017). Fidelity of implementation: An overlooked yet critical construct to establish effectiveness of evidence-based instructional practices. CBE—Life Sciences Education 16(1), rm1. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Stanton, J. D., Dye, K. M., Johnson, M. S. (2019). Knowledge of learning makes a difference: A comparison of metacognition in introductory and senior-level biology students. CBE—Life Sciences Education 18(2), ar24. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Stanton, J. D., Sebesta, A. J., Dunlosky, J. (2021). Fostering metacognition to support student learning and performance. CBE—Life Sciences Education 20(2), fe3. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Stein, M. K., Grover, B. W., Henningsen, M. (1996). Building student capacity for mathematical thinking and reasoning: An analysis of mathematical tasks used in reform classrooms. American Educational Research Journal 33(2), 455–488. [ Google Scholar ] Stein, M. K., Smith, M. S. (1998). Mathematical tasks as a framework for reflection: From research to practice. Mathematics Teaching in the Middle School 3(4), 268–275. [ Google Scholar ] Tekkumru-Kisa, M., Coker, R., Atabas, S. (2022). Learning to teach for promoting student thinking in science classrooms. Teaching and Teacher Education 120, 103869. [ Google Scholar ] Tekkumru-Kisa, M., Schunn, C., Stein, M. K., Reynolds, B. (2019). Change in thinking demands for students across the phases of a science task: An exploratory study. Research in Science Education 49, 859–883. [ Google Scholar ] Tekkumru-Kisa, M., Stein, M. K., Doyle, W. (2020). Theory and research on tasks revisited: Task as a context for students’ thinking in the era of ambitious reforms in mathematics and science. Educational Researcher 49(8), 606–617. [ Google Scholar ] Theobald, E. J., Hill, M. J., Tran, E., Agrawal, S., Arroyo, E. N., Behling, S., … Freeman, S. (2020). Active learning narrows achievement gaps for underrepresented students in undergraduate science, technology, engineering, and math. Proceedings of the National Academy of Sciences 117(12), 6476–6483. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Thurn, C. M., Edelsbrunner, P. A., Berkowitz, M., Deiglmayr, A., Schalk, L. (2023). Questioning central assumptions of the ICAP framework. NPJ Science of Learning 8(1), 49. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Turpen, C., Finkelstein, N. D. (2009). Not all interactive engagement is the same: Variations in physics professors’ implementation of peer instruction. Physical Review Special Topics—Physics Education Research 5(2), 020101. [ Google Scholar ] Vickrey, T., Rosploch, K., Rahmanian, R., Pilarz, M., Stains, M. (2015). Research-based implementation of peer instruction: A literature review. CBE—Life Sciences Education, 14(1), es3. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Wang, M. C., Haertel, G. D., Walberg, H. J. (1990). What influences learning? A content analysis of review literature. Journal of Educational Research 84(1), 30–43. [ Google Scholar ] Waugh, A. H., Green, K. E., Andrews, T. C. (2025). How do early-career biology faculty develop pedagogical content knowledge? Exploring variation and longitudinal development. CBE—Life Sciences Education 24(1), ar13. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials cbe-25-ar11-s001.pdf (520.5KB, pdf) cbe-25-ar11-s001.pdf (520.5KB, pdf) Articles from CBE Life Sciences Education are provided here courtesy of American Society for Cell Biology ACTIONS View on publisher site PDF (709.5 KB) Cite Collections Permalink RESOURCES Similar articles Cited by other articles Links to NCBI Databases Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• A Case Study of a University Induction Program: Coaching Early Career Teachers#201161
• A CROSS-CULTURAL COMPARISON OF PRE-SERVICE MATHEMATICS TEACHER PREPARATION APPROACHES IN JAMAICA AND HISTORICAL BLACK COLLEGES AND UNIVERSITIES IN THE SOUTHEAST UNITED STATES#226448
• Professional Attributes and Teaching Performance: Toward Continuous Improvement in Private Basic Education#402349
• Which Knowledge Do Pre‐Service Teachers Need to Meaningfully Engage in Modeling?#629246
• THE ROLE OF EXPERIENCE IN LEARNING TEACHING PROFESSION: A SYSTEMATIC LITERATURE REVIEW#482451