Learning, and assessing, collaboratively: Group Exams

by Stephanie Chasteen on December 29, 2014

I am one of many who are convinced that people learn better in collaboration with others.  However, there’s always this somewhat disturbing schizophrenia when it comes to assessment — we spend all this time emphasizing group work and collaboration, but come exam time — it’s everyone for him or herself.

So I was very excited to hear about the idea of group exams.  In a group exam, students complete the exam individually (creating a desirable level of individual accountability), and then re-take the exam in a group.

Figure courtesy of Joss Ives, UBC Figure courtesy of Joss Ives, UBC


The exam may be exactly the same, or (for time purposes) just a subset of the questions on the original.  The rationale is that this gives students a chance to review and discuss their answers with each other, while the questions are still fresh in their mind and they are highly motivated to understand their mistakes — rather than a week later when the stress of the exam has faded.  Students can get some level of partial credit for their group answers.

I’ve tried this once, and really liked it — students responded positively to it, and I saw good conversations during the group portion of the exam.

However, I did find it logistically challenging — and now that I see recommendations from the Carl Wieman Science Education Initiative on best practices, I have ideas on how to make them work better — such as making shorter exams, having groups collaborate on a single group exam (whereas I had students correct their own exams), and consider slightly modifying the group portion of the exam.   The group at British Columbia didn’t invent group exams — this is an idea that’s been recycled a few times — but I get the sense that they’ve been somewhat surprised by the level of interest in the approach as they’ve published on their work.  I think this is an idea whose time is ripe — we need to find models of assessment that are more aligned with our instructional approaches.

Joss Ives of the University of British Columbia has been using group exams in his classes for several years, and recently conducted a study to see whether this group exam format actually helps students learn.  Several previous studies had suggested that students do better on the group exam, when the same questions are used on the group portion (see his poster on this topic for references).  But it is difficult to interpret that data — does higher success on the group portion of the test really suggest that students learned more?  Dr. Ives used a slightly more sophisticated approach, by (a) designing re-test questions that were similar, but not the same, as the original test questions, and (b) controlling for other variables, such as student ability and question difficulty.

The results were somewhat modest — student learning was improved for the treatment (i.e., questions which were included on group exams, versus questions which were only included on individual exams) at 1-2 weeks after the group exam, but not for questions 6-7 weeks after the group exam.  These results are somewhat preliminary, however — repetition with other classes, and other questions, may help to determine whether group exams really do improve learning, and Dr. Ives hopes to repeat it in the upcoming year.

But as he says himself, regardless of learning gains, group exams offer many benefits — including better alignment between instruction and assessment, leading to better student buy-in of interactive techniques; increased motivation and enjoyment, and lower failure rates.

Other resources:


{ 1 comment }

Bill Goffe December 29, 2014 at 9:23 pm

I’ve used group quizzes as part of “Team-Based Learning” and I can attest to the engagement it fosters. It is quite amazing to watch students working together and it very much reminds me of the adage that teaching is all about setting up an environment where students learn and not where the teacher simply conveys information to students.

But, I do have a question on how to implement it in a more typical class. I presume that students select their own groups? I wonder a bit about this as I suspect that some groups will be uneven in terms of skills — won’t some tend to have more capable students and others less capable ones? Outside the classroom it seems that students tend to associate with students like themselves and I’d guess that they do in the classroom as well.

In TBL, teams are formed by the instructor and one goal is to have them roughly equal in abilities.

Comments on this entry are closed.

Previous post:

Next post: