Teaching statistics with Peer Instruction and open-source resources

Mine Çetinkaya-Rundel, Assistant Professor of the Practice, just finished teaching her first course at Duke, Statistics 101: Data Analysis and Statistical Inference to about 70 students. She used clickers and Peer Instruction, an open source text, online quizzing and Sakai.

For Peer Instruction (description and evidence for effectiveness of Peer Instruction PDF), students were asked to respond to questions in class using an iClicker. If fewer than 70% of students get the correct answer, students are asked to turn to each other and convince each other that their answers are correct. After 2 to 3 minutes of discussion, students are asked to respond again and most often they converge to the right answer. Dr. Çetinkaya-Rundel finishes with an explanation and answers any remaining questions. She says:

With this approach, students not only interact with me through the use of clickers, but also with each other through small group discussions. In a large lecture it’s always difficult to assess students’ understanding of the material and clickers are a great remedy to this problem. Plus, the feedback is immediate and transparent, meaning that both the professor and the students get to assess how easy or difficult a time students are having with particular concepts.

Students were motivated to participate using clickers because their answers count for part of their grade (students need to answer at least one question right throughout each lecture to get credit). Once the students registered their iClickers, Dr. Çetinkaya-Rundel found that clicker data could be easily uploaded to Sakai to track student grades.

In addition, to increase discussion among students, she may ask a question where there isn’t one, or any correct answers.

Dr. Çetinkaya-Rundel also uses clickers to collect data from students. Survey data can be collected anonymously, which allows her to ask opinion questions on current news or hot topics. Then the class analyzes real and relevant data as they learn a new statistical technique. In addition, Dr. Çetinkaya-Rundel also uses clickers to collect data for class activities such as those designed to convey ideas of randomness. She says:

Class activities are often tricky to implement in a large lecture due to the amount of time required to collect data, but clickers make this process quick and easy, leaving more time for actual discussion of results and findings.

Students also completed weekly online quizzes to give them immediate feedback on their progress in the course.

Students had a choice of using the assigned text (Dr. Çetinkaya-Rundel is one of the co-authors) online for free, or purchasing a copy for under $10. Dr. Çetinkaya-Rundel observed that most students do both, using the text in whatever form is most convenient at the time. Students also use the free, open source statistical software RStudio.

Next semester, Dr. Çetinkaya-Rundel plans to include even more student interaction, because she observed that students learn more when they are actually solving problems.