February 23rd, 2011

Group Exams and Quizzes: The Benefits of Student Collaboration

By:

Have you ever used any sort of group testing activity? The approach is not without benefits. Most students find exams enormously stressful and abundant research documents that high levels of test anxiety can compromise performance. Said more bluntly, students can know the information, but be so anxious they can’t summon it for the exam. Letting students work together on test questions reduces that anxiety considerably. It could be a case of “misery loves company” or the “two heads are better than one” scenario.

Group exams can be structured so that they force students to collaborate. Maybe force is too strong but in competitive learning environments—like those where students are graded on a curve—students can be hurt by helping each other. Because a lot of education emphasizes competition, students are slow to adjust in environments that value cooperation. They won’t offer help unless there are benefits from doing so or risks if they don’t. For example, if everyone in a study group gets a C on the exam, everyone gets a set number of bonus points. If everyone gets a B, they’re awarded even more points, and so on. Or, you can use quiz groups (I’d assemble them with a range of abilities). Every group member is assigned a color and on quiz dates, a color is selected. That student takes the quiz and everyone in the group (present that day in class) gets that grade.

When they work together on exam questions, students talk content, with an intensity and passion that makes teachers (at least this one) quiver with excitement. I was never very successful getting students to engage in debate—to really disagree and argue with each other. But when it’s an exam question, the gloves come off. This is the reason why groups make better decisions than individuals—they deliberate, consider alternatives, state and restate positions, use persuasion, ask questions.

Designing your group exam
Faculty are justifiably wary of group exams because they don’t want students who haven’t learned the material to benefit from others in the group who have. That problem can be avoided by how the group exam activity is designed. In my class, students took the exam individually, turned that exam in and then in their study group were given one clean copy of the exam which they completed together. I first graded all the individual exams. If a student got less than 50%, they failed and got no bonus points. Then I calculated an average score based on the individual exam results of the group members who had passed the exam. Next I graded the group exam. If it was higher than the individual average, I added the difference to each individual score.

The collaboration doesn’t have to occur over the actual exam. It can be a debrief activity. In this case students are convened in groups the day after the exam. They get a copy of the exam and half the time to complete it. If they get a perfect score, everybody in the group gets a specified number of bonus points added to their score.

Group exam experiences can be configured in many different ways. If you don’t want to experiment with an exam, start with a quiz. If you don’t want students sharing a score, give them 15 minutes to work on the quiz individually, then put them in groups of four and give them 5 minutes to discuss questions with each other. Then give them 5 minutes to revisit their own quiz. Strategies like this are also effective at getting students to come to class prepared. Do they want to sit with a group of their peers and admit they don’t know any of the answers? You can make it clear to students that group members don’t have any obligation to help members who are unprepared.

If you’ve used any collaborative exam or quiz strategies, please share your experiences in the comment box below. Questions and comments are also welcome.

Credit for some of the ideas just described belongs to this author: Burrowes, P. A. (2003). A student-centered approach to teaching general biology that really works: Lord’s constructivist model put to a test. The American Biology Teacher, 65 (7), 491-502.