L.E.A.P. – Literature Review of Best Practices in College Physics and Best Practices for Women in College Physics

Kristy Martinez and Margaret Eisenhart

January, 2004

Copyright 2004 by Patricia Rankin and Joyce Nielsen, All rights reserved.

Download Entire Article as PDF

Main Article | Figure 1: List of Articles Read | Figure 2: Review of Evidence-Based Studies (PDF)

Background

Dr. Margaret Eisenhart from the School of Education began a review of research as part of the outreach component of the LEAP project in October, 2002. At that time, Dr. Eisenhart enlisted the assistance of P.h.D. student Kristy Martinez. Our review focused on two components of college women’s experience within the sciences, primarily physics. We searched for evidence-based studies of (1) best practices in college teaching of science, and (2) best practices for teaching college women in science. For each study, we distinguished them by research question, treatment, outcome measures, research design, and actual outcomes. This report synthesizes our findings.

Figure 2 includes the results of each review of each study.

We originally planned to extend the review to best practices in high school science and for high school girls. However, time did not permit us to complete this part.

Best Practices for Teaching College Physics

After searching the literature for “best practices” for teaching college physics since 1997 (the past 5 years) and reviewing the reference lists from this literature, we found twenty relevant articles. Figure 1 includes the full list of relevant articles. Of the 20 articles, only seven qualified as empirical studies of best teaching practices in college physics. The evidence provided in these seven articles suggests that utilizing traditional teaching methods (teacher-directed lecture and problem-solving recitations) in introductory physics courses does not produce desired student outcomes, e.g., students’ conceptual understanding of physics concepts or positive attitudes toward physics and science. Put another way: Traditional teaching methods do not seem to be the best practice for consistently producing these desired outcomes. The studies describe alternative instructional practices and assess their results. The comparisons of traditional and alternative instructional methods suggest that the desired outcomes may be better achieved when students are active participants in class, when instructors have regular, quick ways of assess students’ understanding (flash cards, clickers) during class, and when instructors and students have in-class opportunities to discuss students’ understandings and difficulties. Note, however, that the positive results on desired outcomes come primarily from pre- and post-test comparisons of classes using traditional vs. alternative instructional methods. True control groups have usually not been established, thus the Hawthorne effect and other threats to validity remain unaccounted for. These methodological weaknesses limit the validity and generalizability of the results.
The first study, of a program called Peer Instruction (Crouch & Mazur, 2001), took place at Harvard University between 1990 and 2000. The study compared traditional physics courses with an average of 125 students per class with peer instruction courses with an average of 172 students. The Peer Instruction approach replaces the usual lecture component of instruction with mini lectures followed by intervals of student interaction. In Peer Instruction, introductory physics content was divided into a series of short presentations, each focusing on a central point and followed by a related conceptual question. Students were given time to formulate their individual answers and report their answer to their instructor. Students then discussed their answers with others sitting around them. The instructor then called an end to the discussion and polled the students for their individual answers. The researchers found that students who participated in peer instruction made greater gains on tests measuring conceptual understanding and problem solving than students who participated in traditional courses (Crouch and Mazur, 2001).

The second study, of a program called Fully Interactive Physics Lecture (Meltzer & Manivannan, 2001) and described as a variant of Peer Instruction, was conducted in 997-2002 at 3 universities. The study consisted of 331 students from Southeastern Louisiana University, the University of Virginia, and Southwest Missouri State University. Students participated in class using flash cards to indicate understanding during the lecture. Surveys showed that participating students react favorably to the peer instruction methods. In addition, pre-test to post-test gains were high, tripling those found in national samples using the same tests.

The third study, of a program called Traditional Problem Solving (Kim & Pak, 2001), questioned whether a “traditional problem solving” approach improves conceptual understanding. Twenty-seven students, nine females and eighteen males, who were first-year students in the Physics Education Department at Seoul National University participated in the study in 1994. The study compared students who solved (on the average) 1500 physics problems with those who solved only half as many. The study found that the students who solved the increased number of physics problems continued to demonstrate well known difficulties in basic understanding, although their scores on tests were high. Apparently, simply increasing the number of problems solved does not improve conceptual understanding.

The fourth study, of a program called Interactive Engagement Methods (Hake, 1998), took place during 1992 at the University of Indiana. The study involved various methods of making the physics environment more interactive. Interactive engagement activities included hands-on activities that brought immediate feedback followed by discussion with peers and instructors. Test score results demonstrated only low to medium gains as a result of interactive teaching methods, but conceptual understanding scores showed promising improvements. Teacher effects, e.g., familiarity with physics education research, are suggested as an important factor in these outcomes, although they were not measured in this study.

The fifth study, of a program called the Maryland Physics Expectation Survey (Redish et al., 1997), examined students’ attitudes, beliefs, and assumptions towards physics before and after taking an introductory calculus-based physics course. The survey group consisted of 1500 students from various colleges. The study was conducted in 1997 at the University of Maryland, University of Minnesota, Ohio State University, and Dickenson College and found that taking this particular introductory physics course lowered rather than raised students’ expectations of science.

The sixth study, of a program called Audience Based Feedback (Poulis et al., 1996), examined the effects of regular audience (student) feedback (20 minutes per class) on end-of-course pass rate, achievement, and attitudes. The study conducted between 1979 and 1992, at Eidhoven University of Technology, surveyed 5391 students. Feedback focused on students’ comprehension, ability to apply principles, and pace of instruction. The results were that students in the audience based feedback courses had a significantly higher pass rate, demonstrated less variability in achievement, and had more positive attitudes than students in traditional lecture classes.
The final study examining best teaching practices in physics, done between 1993 and 1995, explored the use of active-engagement, microcomputer-based labs as a substitute for traditional problem-solving recitation sessions in introductory calculus-based mechanics classes (Redish et al., 1996). The survey group consisted of 553 students at the University of Maryland. Students were given one-hour active-engagement tutorials using microcomputer-based laboratory equipment. Tutorials included concepts such as instantaneous velocity and Newton’s third law. Students responded to the tutorials by answering multiple choice as well as free-response questions. Students in the microcomputer based labs produced better overall test gains than those in regular recitations.

Best Teaching Practices for College Women in Physics

We then turned to best practices for teaching women in college physics courses. Although the number of studies in this category is very small, similar patterns emerge in their results.

We found only two evidence-based studies of best practices for women in physics. The first study, of a program called Workshop Physics (Laws et al., 1999), consisted of questionnaires and interviews of 46 students in Introductory Physics courses between 1989 and 1990 at Dickinson College. Attitude surveys also were given to approximately 2800 students at 14 other U.S. colleges or universities in 1990. The study compared women’s and men’s attitudes and achievement after participating in workshop physics. Students were taught in a four-part sequence utilizing microcomputers and various scientific apparatuses. Students began a topic with an examination of their own preconceptions followed by qualitative observations. Students were given time for discussion, followed by the instructor’s assistance with definitions and mathematical theories. Discussion ended with qualitative experimentation to verify mathematical theories (Laws et al., 1999). Women were more positive about some aspects of Workshop Physics than men (e.g., the learning lab), while average grades and the proportion of students likely to major in physics after taking Workshop Physics were roughly the same for women and men.
The final study examined best practices for teaching women and at-risk students in physics. The study compared two non-equivalent groups—one group of at-risk students, mainly women (n=300), taking the alternative Extended General Physics, and one group of regular students taking traditional General Physics (n=5000) (Etkina et al., 1999). The study was conducted at Rutgers University between 1992 and 1997. Extended General Physics included active participation, cooperation among students, and activities designed to be fun as well as relevant. These features were not present in the traditional General Physics classes. The results were that the at-risk students in Extended Physics had higher course retention rates, rated their instruction more highly, and scored higher on common instruments than the students in General Physics. Another result was that the cost per successful student was roughly the same, despite the fact that Extended Physics is much more labor-intensive for colleges and universities.

Conclusion

In summary, our review suggests that when desired outcomes for college students (including women) are conceptual understanding in physics and positive attitudes toward physics and science, these outcomes are better achieved using alternative instructional methods than with traditional instructional methods. The alternative instructional methods that show promise are various activities that allow students to be active class participants, procedures or devices that give instructors quick ways to assess their students’ understanding during class, in-class opportunities to discuss students’ understandings and difficulties, and activities specially designed to be fun, challenging, and relevant. However, the small number of studies and their methodological weaknesses limit the validity and generalizability of the results and point to the need for more controlled studies of effective college physics instruction.