October 10th, 2012

Peer Assessment is Not an Elixir for All Group Work Challenges


Here’s an empirical result I would not have predicted. It emerged in a meta-analysis of research on group learning activities as reported in 32 studies published in Teaching of Psychology between 1974 and 2011. I’ll be doing a detailed highlight of the entire article in an upcoming issue of The Teaching Professor. But the finding that surprised me involved the use of peer assessment within groups: “Our hypothesis of better learning outcomes with peer assessment was not supported. In fact, the data suggest that the opposite pattern may exist.” (p. 164)

Having group members assess each other’s contributions is a common practice. Many teachers use it to encourage students to accept their responsibilities within groups. It’s hard to understand why an approach that aims to get students contributing to the group’s success wouldn’t positively affect learning outcomes. Perhaps the problem is as these authors suggest—group members are more focused on being evaluated by their peers and less committed to learning the material.

Maybe our primary motivation for using peer assessment shouldn’t be as a mechanism for dealing with students who aren’t contributing as they should. It might be better to use peer assessment to teach principles of constructive feedback. Most professionals are expected to offer feedback and otherwise evaluate the work of peers. It’s a complicated skill that requires no small amount of interpersonal finesse. To do it well, students need opportunities to practice delivering this kind of feedback.

I’m not sure that when the peer assessment involves filling out some sort of rating instrument at the conclusion of a project it does much to develop feedback skills. Group members need to discuss rating criteria before they start working together. They need to incorporate some sort of formative assessment as they work so that individuals have the chance to modify their behavior. Moreover, assessment feedback at the end of the project is too late to change what happened in the group. Those discussions need to occur as the group work proceeds, not after the fact.

We’ve still got the problem of individuals who don’t carry their weight in the group. There is evidence (several studies are cited in this meta-analysis) that peer assessment does effectively address that problem. It does motivate members to meet their commitments to the group, but peer assessment doesn’t improve, and maybe even compromises, learning outcomes (as this meta-analysis suggests). Is that too high a price to pay?

Groups can be empowered to deal with nonproductive members in ways other than peer assessment, and some of the approaches are pretty straightforward. It’s difficult to be a silent member in a group if others regularly and directly ask for your opinion and input. Groups can assign preliminary tasks and set deadlines, and use that process to see who does and doesn’t deliver work on time. Groups can put pressure on individual members. Doing so constructively is part of learning how to help groups function productively. All too often when a student in the group isn’t contributing, the group does nothing, and ignoring the problem makes it just that much easier to continue doing nothing.

I don’t think the researchers’ finding that peer assessment didn’t improve learning outcomes is a mandate to abandon use of peer assessment in groups. Rather it challenges us to explore the use of peer assessment. How do we make sure it contributes positively to overall group functioning and encourages individuals to accept their responsibilities? How can it be used so that it develops the feedback skills of individual members? And finally, are there approaches to peer assessment that do promote, not compromise, the learning experiences of group members?

As I’ve been working on this post, I’ve been thinking that peer assessment beautifully illustrates a basic fact of teaching. It looks like the problem is simple; students are not doing their fair share of the work in a group. It looks like the solution is easy; have students assess each other’s contributions. But this unexpected finding yet again confirms that most (I’m tempted to write all) teaching problems and solution are not simple or easily fixed.

Reference: Tomcho, T. J. and Foeis, R. (2012). Meta-analysis of group learning activities: Empirically based teaching recommendations. Teaching of Psychology, 39 (3), 159-169.