Peer Instruction Research

In classical Peer Instruction, the instructor poses a question to the class on a topic just discussed (or that the students read about before class) and polls them for their individual responses.  Based on the results s/he chooses to move on to the next topic or not. If more attention is needed, students are asked to discuss the reasoning behind their responses with their immediate neighbors; after a few minutes the instructor re-polls.  Variations include the instructor showing the students the polling results after the first round or not, choosing whether to ask follow-up questions for polling/discussion, and how to cover the topic prior to polling.

Instructors[1] who implemented PI in this sense were initially polling the class in order to ascertain whether students were learning in lecture-style classes; the "turn to your neighbor" aspect was a follow-on aimed at overcoming student confusion in


Find Related Activities



specific cases. Thus the research focuses on learning gains, both after a given poll/discussion/re-poll episode and for entire courses.  Studies in STEM fields show that PI can greatly increase student comprehension of conceptual material[2] even when the questions asked are not themselves purely conceptual[3].  While it is most prevalent in STEM fields, PI has been implemented in other fields [for example (4), and see also (5) and (6)].

Even though the technique has come to be associated with specific tech (clickers and more sophisticated response systems), the main result from the research is that it is the students' interactive struggling with concepts that produces learning gains, which are likely to occur even when two students discussing together both initially had an incorrect response[5]. A low-tech implemention involves simply provide students with a sheet of paper with each letter "A B C D" (representing response choices) in a different corner of the page, in different colors, so they can fold it to show only one letter and so vote with it for easy, if appproximate visual collation by the instructor.  For gathering data for research however a more robust system (clickers, Learning Catalytics, etc.) would be required.



- J. Girash, Harvard University



[1] Crouch, C. H.; Mazur, E. Peer Instruction: Ten years of experience and results. Am. J. Phys. 2001, 69, 970.
[2] Fagen, A. P.; Crouch, C. H.; Mazur, E. Peer instruction: Results from a range of classrooms. The Physics Teacher 2002, 40, 206.
[3] Miller, R. L.; Santana-Vega, E.; Terrell, M. S. Can good questions and peer discussion improve calculus instruction? Problems, Resources,
and Issues in Mathematics Undergraduate Studies 2006, 16, 193-203.
[4] Butchart, S.; Handfield, T.; Restall, G., 2009.  "Using Peer Instruction to teach Philosophy, Logic and Critical Thinking". Teaching Philosophy, 32:1, 1-40.
[5] Turn to Your Neighbor ("Official PI blog") blog.peerinstruction.net
[6] Agile Learning blog derekbruff.org/teachingwithcrs
[7] for example, learningcatalytics.com
[8] Smith, M. K.; Wood, W. B.; Adams, W. K.; Wieman, C.; Knight, J. K.; Guild, N.; Su, T. T. Why peer discussion improves student performance on in-class concept questions. Science 2009, 323, 122-124.

Additional resources: