Skip to main content

research & teaching

Who Gets Helped?

The Opportunity Structure of the College Physics Classroom, Peer Instruction, and Perceptions of Help Seeking

Journal of College Science Teaching—November/December 2019 (Volume 49, Issue 2)

By Michael Brown and Robert M. DeMonbrun

Although peer interaction in the college classroom may benefit some students by developing peer networks for future collaborations, prior research has suggested that these relationships most often benefit students from overrepresented communities, while minoritized student populations may gain less from these experiences. In this article, we explore the opportunity structure in a physics course to identify differences in perception of classroom community among students. Using multinomial logistic regression (n = 551), initial findings suggest that students who collaborated had higher odds of identifying their relationships as equally beneficial. Among those who perceived differences in their relationships, White students were more likely to report they received help, and minoritzed students were more likely to report providing help, especially when they worked in homophilous pairs.


The undergraduate physics curriculum is organized around a disciplinary “logic of collaboration” where students must work together to successfully move through the curriculum (Nespor, 1994, p. 40). Because of the tightly sequenced curriculum, students in physics traditionally find that their social and academic networks become entangled (Nespor, 1994) to such an extent that students with less overlap between their social and academic worlds have a harder time persisting in the major (Forsman, Linder, Moll, Fraser, & Anderson, 2014). Given the logic of collaboration that predominates in physics, it is perhaps unsurprising that nearly a quarter of American undergraduate physics instructors have adopted the active learning strategy called peer instruction (PI; Henderson & Dancy, 2009). PI requires students to interact to solve problems during class time and is intended to facilitate the development of peer networks that students will draw on throughout their undergraduate careers (Mazur, 2009).

Substantial evidence suggests that PI approaches improve conceptual reasoning and quantitative problem solving in calculus-based introductory physics courses (Crouch & Mazur, 2001) as well as improved course retention (Lasry, Mazur, & Watkins, 2008). Yet, the pedagogical assumptions that inform PI potentially underestimate the distribution of opportunity to participate in active learning and form peer collaborations in college classrooms. The structure of educational opportunity on campus and within courses shapes who gets to access what resources during class time and on campus (Tinto, 1997). PI instructional practices may ignore potential differences among students within the educational opportunity structure, presuming equitable participation and access to peers who can provide social and academic resources. Although all students may, on the average, benefit from PI, social forces exogenous to the classroom inevitably shape who students choose to collaborate with, leaving students who are minoritized with differential access to peer collaborators. 

Related literature

PI refers to an instructional approach developed by Eric Mazur and his colleagues at Harvard (Mazur, 2009), which “engages students during class through activities that require each student to apply the core concepts being presented, and then to explain those concepts to their fellow students” (Crouch & Mazur, 2001, p. 970). The PI model proposed by Mazur (2009) shifts the focus of the lecture hour from information transfer to knowledge construction. Students review material before attending class and spend class time answering conceptual problems. Then students are presented with brief presentations on a topic and a series of multiple-choice questions. The instructional team (a lead instructor and teaching assistants) circulates around the room and provides feedback to students (Mazur, 2009). Next, students are polled again, and the instructor explains the correct solution. Thus, students are encouraged time and again to collaborate and dialogue with their peers in the course, fostering a network of social connections between students as part of the teaching and learning process. 

As such, we might expect that using PI will in some way shape emergent classroom peer networks. Classrooms where instructors encourage interaction foster the development of social ties between students (Carolan, 2013). Although very little empirical literature examines how student collaborations emerge in postsecondary classrooms, a substantial body of literature suggests that social ties—especially friendship ties—between college students are generally guided by homophily, or similarity along dimensions of shared identity and/or experiences (Biancani & McFarland, 2013; McPherson, Smith-Lovin, & Cook, 2001). Whether similar factors shape in-classroom collaborations is unclear, but as faculty adopt instructional practices that engage students through peer interaction, an understanding of how student collaborations emerge is needed. 

Concerns about students’ perceptions of their interactions during peer instruction are warranted in part because the composition of groups potentially has an impact on how minoritized students perform in science and engineering courses (e.g., Dasgupta, Scircle, & Hunsinger, 2015; McCullough, 2002; Stout, Dasgupta, Hunsinger, & McManus, 2011). In fact, the use of PI raises an important concern and a potential unintended consequence: Given the underrepresentation of women and students of color in physics courses (Ivie & Ray, 2005; Kanim, 2015), an instructional technique that relies on collaboration to facilitate learning might produce unequal outcomes for students who are less likely to be included in a network of collaboration shaped by homophily.

A sorting process occurs in all classrooms that reflects inequities in access to power and the resources that flow from relationships (M. Brown, 2017). Students make conscious decisions about whom to include and exclude in ways that conform to their preferences and biases (McCabe, 2016). If emergent classroom networks conform with other research on interpersonal social networks (e.g., Goodreau, Kitts, & Morris, 2009) and no effort is made to disrupt existing social structures, they will most likely form in ways that exclude some students and privilege others. For example, women and students of color were more likely to be excluded in in-class and out-of-class study groups in large STEM (science, technology, engineering, and mathematics) courses (e.g., Callahan, 2008; Freeman, Theobald, Crow, & Wenderoth, 2017). 

In this study, we examine peer instruction in three sections of an introductory physics lecture course with varying levels of PI implementation to identify whether students perceive the opportunity structure of the course in substantially different ways or if systemic demographic inequity in undergraduate physics education persists across pedagogical approaches. In doing so, we explore how students perceive the opportunity structure in PI-intensive and lecture-intensive sections as well as the relationship of students’ perceptions of the opportunity structure to academic performance.

Research questions

Given the initial findings from the prior literature, we ask the following:

  • RQ1: What factors are related to significant changes in students’ perceptions of their peer interactions in an introductory physics course using PI?
  • RQ1A: What factors are related to significant changes in students’ perception that they mostly received help in the course, when compared with students who reported equally beneficial relationships?
  • RQ1B: What factors are related to significant changes in students’ perception that they mostly provided help in the course, when compared with students who reported equally beneficial relationships?
  • RQ2: Are students’ perceptions of the classroom opportunity structure significantly related to differences in performance at the end of an introductory physics course using PI?

Research context and sample

The context for this study is an introductory physics course at a university in the American Midwest. According to the 2010 Carnegie Classification, the institution is a high undergraduate enrollment, Doctoral/Research University–Extensive. The majority of undergraduate students live on campus. 

The course is the first half of a two-semester introductory sequence to concepts in mechanics. Students attend a 1-hour lecture three times a week as well as a weekly 3-hour lab section and a 1-hour discussion section. Different instructors lead each section of the lecture. Graduate assistants administer lab sections. The three lecture instructors were all White cisgender men. 

On the basis of daily observations of the course section by the first author, we characterized each section of the lecture course (and each instructor) according to their frequency of use of the PI approach. In two of the sections (which we refer to as active PI sections), the instructors used PI activities throughout the lecture section. They lectured rarely, provided students upward of 10 iClicker questions in a session, and regularly encouraged students to interact and solve problems collaboratively during class time. The other instructor relied more often on a traditional lecture approach with PI activities occasionally incorporated. In this section, which we refer to as the direct lecture section, the instructor engaged students in at least one PI activity during every class. However, the balance of class time never tipped over into PI activities compared with the time spent lecturing. Two of the instructors were tenured full professors (one in an active section, one who spent the majority of time lecturing). The other active PI section was taught by a full-time senior lecturer. This course is a requirement for popular major programs in physics, chemistry, biochemistry, and engineering. 

The sample for this study is composed of students enrolled in an introductory undergraduate physics course designed for engineers. Of the 511 students in the course, 450 completed both of the surveys we administered. Students received extra credit points on an exam for completing the surveys. Descriptive statistics for the study sample are presented in Table 1. The sample was generally representative of institutional enrollment, except for women who were significantly underrepresented (accounting for less than 30% of students in the course and the sample).

TABLE 1. Course demographics by class section (N = 551).


Class A
n = 151)

Class B
n = 176)

Class C
n = 221)


69% (106)

73% (129)

70% (156)


31% (47)

27% (47) 

30% (65)


25% (38)

31% (54)

26% (58)


1% (2)

2% (4)

1.4% (3)


7% (11)

4.5% (8)

7.2% (16)


4% (6)

4% (7)

3.6% (8)


0.7% (1)

0.5% (1)


Not indicated

4.6% (7)

1.1% (2)

1.9% (4)


58% (88)

56% (100)

59% (132)


9.9% (15)

10.2% (18)

5.8% (13)

Survey response rate





Literature, arts, & sciences

21% (32)

32% (56)

23% (52)


79% (121)

68% (120)

76% (169)


Note. AAPI = Asian American/Pacific Islander. NA/NH = Native American/Native Hawaiian.

Data collection 

Data from students were collected through the administration of two surveys. Both surveys were administered on the internet to students’ institutional e-mail address. The first survey was administered starting the fifth week of the course in the lead up to the first exam. The second survey was administered during the 12th week leading up to the third exam. There were four exams in total in the course. Students were provided extra credit for participating in the study. As part of the informed consent process, students granted the researcher access to their institutional data and grades in the course. 

On the survey, students were asked who they worked with in the course. This was completed using an aided recall roster, where students would start to enter the name of a peer and an autogenerated list of potential matches would appear. Students could also enter a peer’s nickname or their email address. Students could select “name unknown” and still provide demographic details about their study partner and answer questions about their relationship; however, no students used this option. This aided-recall approach was developed through two pilot studies, where students’ survey reports were cross-referenced during class time. In both pilot studies, students’ survey responses were reliably indicative of their actual study partners. 

Students were also asked if as part of their study partner relationship they generally perceived that they received help, they provided help, or the relationship was equally beneficial. Students were asked, in general, while completing homework and preparing for assessments, if they were able to access help when they needed it. This second measure was used to determine the analytical sample, as students that reported that they never needed help in the course were not included in the final analysis. The final analytical sample was 439 students. 


The surveys contained items drawn from Eccles’s (2005) Expectancy Value (EV) scale and Rovai’s (2002) Classroom School Community Index (CSCI). The Eccles’s EV scale used in this study was adapted and validated for STEM courses in higher education by Perez, Cromley, and Kaplan (2014). The scale measures components of achievement motivation such as students’ expectations for success, their subjective valuing of course tasks, and the opportunity cost they perceived was required for success in the course (Perez et al., 2014). The CSCI contains a scale that measures sense of community (SOC) in a classroom and another scale measuring SOC on campus (for items and distribution of responses, see the appendix, available at 

Per Garcia, López, and Vélez’s (2018) suggestions for critical quantitative studies in education research, the first author also collected data about instruction from daily observations in the lecture sections of the course to better characterize the classroom environments. This allowed us to document variations among the three instructors in instructional approach, how they framed in-class peer collaboration activities, and how—if at all—they provided students guidelines for providing and seeking help. The first author also conducted periodic observations of each lab section and of exam preparation sessions. 

Methods of analysis

To answer our first research question, we estimated a multinomial logistic regression model to identify significant relationships among socio-demographics, the focus of students’ collaborative study behavior, sources of students’ achievement motivation in the course, and their sense of connection to community. 

The outcome of interest in the multinomial logistic regression model is whether students perceived that their PI study partnerships were (a) equally beneficial, (b) mostly providing help to a partner, or (c) mostly receiving help. In the analysis, we used “equally beneficial” as the reference group; thus, all coefficients are in reference to equally beneficial relationships. We also included controls for students’ race, gender, undergraduate college, whether they were friends with their study partner before the course, and whether they worked together on (a) in-class or (b) out-of-class activities or (c) if they engaged in both. Although all course sections used peer instruction during class time, two of the sections used it consistently throughout the class period, whereas the third used the approach sparingly. We control for which section a student was in; the two sections that used PI consistently might encourage more interaction (which could result in different outcomes), compared with the section where the majority of the time was spent on direct lecture instruction. We also include a covariate for whether students in a pair matched on different demographics (race/ethnicity, gender, course section). Finally, we include the expectancy value and sense of connection scales described previously. 

To answer our second research question, we estimated an ordinary least squares (OLS) regression model, where the outcome was students’ final performance in the course on a scale from 0 to 100 points. We also controlled for students’ demographics; their EVT and CSCI scores; whether they worked together in class, out of class, or both; their lecture section; the number of students they reported working with; and the direction of perceived help in the relationships they reported (on average). Both analyses performed well on tests of goodness of fit. 


As indicated in Table 2, across all three sections of the course, about 40% of students who reported a study partner acknowledged that they were friends with their study partner before the course started. The majority of students who reported a study partner were in active PI sections (73%). Students who reported a study partner mostly worked with that individual during class time on in-class activities (63%) compared with students who worked together only on out-of-class activities (16.9%) and students who engaged in both (19.9%). Among these students, most class time was spent on iClicker questions as part of PI (62%). Less than a quarter of students worked on homework assignments together (23.6%), with slightly more getting together to prepare for exams (28%). These initial results suggest that the majority of students’ peer interactions were spent during class time on instructional activities.

TABLE 2. Course type and study activities of study pairs.

Dyad census


Prior friendship


Active learning dyad


Traditional lecture dyad


Study for exams together


Work on iClickers together


Work on homework together


Worked together only in class


Worked together only out of class


Work together in and out of class


About half of students reported working with peers as part of the course. The majority of students (64%) perceived these relationships as equally beneficial. However, there were substantial differences among students in terms of their perceptions of help received and provided (Table 3).
White students had higher odds of reporting they received help from their partners, compared with Asian, Asian American, and Pacific Islander students (odds ratio [OR]: White = 3.85). Students who worked with friends they knew before the course were over four times more likely to report that they received help from their partner (p < .01). Similarly, students who worked together on either in-class academic tasks or out-of-class academic tasks (rather than equally participating in both) had much greater odds of reporting that their partner primarily helped them, compared with students who worked together in and out of class (OR= 13.96, p < .05; OR = 42.59, p < .01, respectively). White students who worked with another White student and women who worked with other women were less likely to report that they helped their partner more, compared with equally supportive study relationships (p < .01). Higher levels of intrinsic interest in the tasks of the course were significantly related to greater odds of reporting help received and smaller odds of help provided when compared to mutual relationships (OR = 2.69, p < .01; OR = 0.32, p < .01). Students with a higher sense of connection to the classroom academic community had greater odds of reporting providing help (OR = 2.4, p < .1).
Increased connection to the social community on campus was also related to greater odds of providing support and smaller odds of receiving support (OR = 3.07, p < .1; OR = 0.23, p < .1). 

TABLE 3. Multinomial logit for perceptions of helping.


Helped me (>)a

Helped my partner (<)a

Minoritized (Black, Latinx, & American Indian)b









College of Engineeringd



Prior friendship



Worked together: in classe



Worked together: out of classf



Active learning class



Same class



Same race: Whiteb



Same race: minoritizedb



Same gender: womenf



Expectancy value (motivation)
















Sense of community

Class: academic community



Class: social community



Campus: academic community



Campus: social community



AIC: 361.0801

aReference group is mutual support. bReference is Asian, Asian American, and Pacific Islander. cWomen. dCollege of Arts & Sciences. eWorked together in and out of class. fMen.

***p > 0.01. **p > 0.05. *p > 0.1.


As shared in Table 4, predicted probabilities by race and gender subgroup analyses reinforce the notion that students perceived their peer interactions in both courses to be equally beneficial (e.g., the average student had a 76% predicted chance of reporting an equally beneficial relationship). Far fewer students perceived imbalances in the direction of help provided with an average predicted probability of 15% for students who perceived receiving more help and 8.3% for students who perceived that their partner provided more help. These trends are relatively consistent with two exceptions: First, underrepresented men, Asian American/Pacific Islander (AAPI) women, and White women and men had higher predicted probabilities of perceiving that they received help (10%, 20%, 23%, 15%), and underrepresented women had nearly equal odds of perceiving that they received help and provided help (11%, 12%). These results suggest subtle differences in students’ perceptions of peer interactions and the opportunity structure of the course by salient identities of gender and race/ethnicity. 

TABLE 4. Average marginal effect by perceptions of helping.


Partner helped me more

I helped my partner more

Helped each other equally

























URM men




URM women




AAPI men




AAPI women




White men




White women




Note. URM = underrepresented minority. AAPI = Asian American/Pacific Islander.


In terms of differences in performance (Table 5), students who perceived that they received help scored nearly seven points higher on their overall academic performance, on average, than students who perceived equally helpful relationships (B = 6.77, p < .01). We also observed significant differences in performance by demographics, a phenomenon in introductory physics courses that has been extensively documented elsewhere (e.g., Matz et al., 2017). Intrinsic interest was significantly related to poorer overall academic performance (B = –2.09, p < .05), whereas perceived competence (B = 1.88, p < .05) was significantly related to higher levels of performance. Students who were in the more traditional PI section (where the balance of instruction leaned toward direct lecture) outperformed their peers in the more active sections (B = 6.32,
p < .001). This is most likely a byproduct of the focus of preparing for the exam in this section and not an indication of the impact of frequency of PI use on the quality of learning in each section. Students who did not work in pairs with preexisting friends performed better than their peers who worked with friends (B = 2.83, p < .05). Instructors might consider the impact of letting students self-select their study partners for PI activities, as our findings suggest students who rely on existing relationships do not benefit at the same rates as their peers who sort on the basis of other factors. 

TABLE 5. Academic performance and perceptions of helping.



Standard error

Intercept (grade)***



Asian, Asian American & Pacific Islandera









Native Americana



Not indicateda









College of Engineeringc***


















Academic community (class)



Academic community (course)



Social community (class)



Social community (course)



Prior friendship**



In-class collaborationd



Out-of-class collaborationd



Lecture sectione***



# of other students worked with



Perception of helping (provided help)f



Perceptions of helping (received help)f***



aReference group is multiracial. bWomen. cCollege of Arts & Sciences. dWorked together in and out of class. eActive learning section. fMutual support.

***p > 0.01. **p > 0.05. *p > 0.1.



This study contributes to the ongoing exploration of how active learning pedagogy may contribute to equity and performance in undergraduate education (e.g., Gourlay, 2015; Zepke, 2015). In this study, instructors treated peer collaboration as a neutral process, something that students had equal opportunity to engage in. But engagement is an emergent, contingent, and restless process (Gourlay, 2015) shaped by factors external to the classroom and campus and structured in ways that reproduce social stratification (e.g., Armstrong & Hamilton, 2013). In general, our findings provide some support that students appear to be able to access help during peer instruction activities in relatively equitable ways. In most cases, students perceived that their interactions were equally beneficial, and we observed that among students by race and gender, students who did not report beneficial relationships were more likely to report receiving help. These findings are promising as they suggest that PI has the potential to disrupt inequitable education opportunity structures.

However, there are some findings that suggest instructors should carefully consider the impact of PI approaches on the distribution of labor in the classroom. Students who were minoritized in the classroom had much higher odds of reporting that they provided help. While students perceived equally beneficial relationships in the aggregate, minoritized students appear to be spending more time and energy providing support to their peers. Instructors should consider the distribution of work among students within the learning community of the classroom when they engage students in active learning strategies like PI. 

Implications for practice

Our results also provide some evidence that PI strategies may not reduce the persistent inequalities in academic performance among underrepresented students (e.g., women, Black, Latinx, and Native American students) that other researchers have observed in introductory college physics courses (Matz et al., 2017). If equity is an instructor’s goal (and we argue that this should be part of the decision-making process in planning academic tasks), instructors should consider to what extent PI addresses that objective. As instructors go about encouraging peer interaction in their course, they should consider culturally affirming practices for active learning (e.g., J. C. Brown, 2017; Kumar, Zusho, & Bondie, 2018). 

When we compare sections where students are more actively engaged with sessions that employ primarily direct lecture, we find that direct instruction may be more effective at preparing students for assessments. This finding is not to suggest that direct instruction is more effective at fostering learning in introductory physics—there is a substantial body of evidence that suggests otherwise (e.g., Hake, 1998; Prince, 2004). Rather, future researchers and instructors might consider the frequency with which PI activities are used during class time. It may be that a healthy balanced mix of instructional activities engages students and provides deeper content knowledge than relying on one approach exclusively. 

Regardless, we hope this research encourages instructors to consider the structure of opportunity for peer interactions in PI classrooms. In this study, instructors allow students to self-select peers, and our results indicate that this approach might actually result in lower levels of performance. Instead, we encourage instructors and researchers to consider the benefits of random assignment in active learning activities. Such an approach has the potential to radically upend the opportunity structure and may encourage more diverse sorting in peer interactions outside of class as well. ■

Michael Brown ( is an assistant professor in the School of Education at Iowa State University in Ames. At the time this article was written, Robert M. DeMonbrun was a doctoral student at the University of Michigan in Ann Arbor.


Armstrong, E. A., & Hamilton, L. T. (2013). Paying for the party. Cambridge, MA: Harvard University Press.

Biancani, S., & McFarland, D. A. (2013). Social networks research in higher education. In M. B. Paulsen (Ed.), Higher education: Handbook of theory and research (pp. 151–215). New York, NY: Springer. 

Brown, J. C. (2017). A metasynthesis of the complementarity of culturally responsive and inquiry-based science education in K–12 settings: Implications for advancing equitable science teaching and learning. Journal of Research in Science Teaching, 54, 1143–1173.

Brown, M. (2017). Interaction and mechanics: Understanding coursework engagement in large science lectures. Unpublished doctoral dissertation, University of Michigan, Ann Arbor.

Callahan, K. M. (2008). Academic-centered peer interactions and retention in undergraduate mathematics programs. Journal of College Student Retention: Research, Theory and Practice, 10, 361–389. 

Carolan, B. V. (2013). Social network analysis and education: Theory, methods & applications. Thousand Oaks, CA: Sage.

Crouch, C. H., & Mazur, E. (2001). Peer instruction: Ten years of experience and results. American Journal of Physics, 69, 970–977. 

Dasgupta, N., Scircle, M. M., & Hunsinger, M. (2015). Female peers in small work groups enhance women’s motivation, verbal participation, and career aspirations in engineering. Proceedings of the National Academy of Sciences, USA, 112, 4988–4993. 

Eccles. J. S. (2005). Subjective task value and the Eccles et al. model of achievement-related choices. In A. J. Elliot & C. S. Dweck (Eds.), Handbook of competence and motivation (pp. 105–121). New York, NY: Guilford Press.

Freeman, S., Theobald, R., Crow, A. J., & Wenderoth, M. P. (2017). Likes attract: Students self-sort in a classroom by gender, demography, and academic characteristics. Active Learning in Higher Education, 18, 1–12. 

Forsman, J., Linder, C., Moll, R., Fraser, D., & Andersson, S. (2014). A new approach to modelling student retention through an application of complexity thinking. Studies in Higher Education39, 68–86.

Garcia, N. M., López, N., & Vélez, V. N. (2018). QuantCrit: Rectifying quantitative methods through critical race theory. Race, Ethnicity, and Education, 21(2). 

Goodreau, S. M., Kitts, J. A., & Morris, M. (2009). Birds of a feather, or friend of a friend? Using exponential random graph models to investigate adolescent social networks. Demography46, 103–125.

Gourlay, L. (2015). ‘Student engagement’ and the tyranny of participation. Teaching in Higher Education, 20, 402–411.

Hake, R. R. (1998). Interactive-engagement versus traditional methods: A six-thousand-student survey of mechanics test data for introductory physics courses. American Journal of Physics, 66, 64–74.

Henderson, C., & Dancy, M. (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). 

Ivie, R., & Ray, K. N. (2005, February). Women in physics and astronomy, 2005 (AIP Publication No. R-430.02). College Park, MD: American Institute of Physics. 

Kanim, S. (2015). Physics Education Research efforts to promote diversity: Challenges and opportunities. Bulletin of the American Physical Society60

Kumar, R., Zusho, A., & Bondie, R. (2018). Weaving cultural relevance and achievement motivation into inclusive classroom cultures. Educational Psychologist, 53(2), 78–96.

Lasry, N., Mazur, E., & Watkins, J. (2008). Peer instruction: From Harvard to the two-year college. American Journal of Physics, 76, 1066–1069. 

Matz, R. L., Koester, B. P., Fiorini, S., Grom, G., Shepard, L., Stangor, C. G., . . . McKay, T. A. (2017). Patterns of gendered performance differences in large introductory courses at five research universities. AERA Open3(4), 2332858417743754.

Mazur, E. (2009). Farewell, lecture? Science, 323(5910), 50–51. 

McCabe, J. M. (2016). Connecting in college: How friendship networks matter for academic and social success. Chicago, IL: University of Chicago Press. 

McCullough, C. (2002). Attracting under-represented groups to engineering and computer science. Population, 20, 1–11.

McPherson, M., Smith-Lovin, L., & Cook, J. M. (2001). Birds of a feather: Homophily in social networks. Annual Review of Sociology, 27, 415–444.

Nespor, J. (1994). Knowledge in motion: Space, time, and curriculum in undergraduate physics and management. New York, NY: Routledge. 

Perez, T., Cromley, J. G., & Kaplan, A. (2014). The role of identity development, values, and costs in college STEM retention. Journal of Educational Psychology, 106(1), 315–329. 

Prince, M. (2004). Does active learning work? A review of the research. Journal of Engineering Education93, 223–231.

Rovai, A. P. (2002). Development of an instrument to measure classroom community. The Internet and Higher Education5, 197–211. 

Stout, J. G., Dasgupta, N., Hunsinger, M., & McManus, M. A. (2011). STEMing the tide: Using ingroup experts to inoculate women’s self-concept in science, technology, engineering, and mathematics (STEM). Journal of Personality and Social Psychology, 100, 255–270.

Tinto, V. (1997). Classrooms as communities: Exploring the educational character of student persistence. The Journal of Higher Education, 68, 599–623.

Zepke, N. (2015). Student engagement research: Thinking beyond the mainstream. Higher Education Research & Development, 34, 1311–1323.

Physics Teaching Strategies

Asset 2