Collaborative Testing Improves Higher-Order Thinking

Three students working on a group exam on a laptop

Most faculty don’t respond enthusiastically to the idea of students doing exam or quiz work together in groups. Nonetheless, the approach is widely used, and the research continues to show significant benefits. Innovative design features like those in the study below answer many faculty objections. The grades can still be measures of individual mastery of the material, and individual scores can count significantly more than the group score.

The study

Mahoney, J. W., & Harris-Reeves, B. (2019). The effects of collaborative testing on higher order thinking: Do the bright get brighter? Active Learning in Higher Education, 20(1), 25–37.

The research questions

  • What are the effects of collaborative testing on student exam performance? Can previous positive effects be further confirmed?
  • Does collaboration improve performance on exam questions that require higher order thinking skills?
  • What were students’ perceptions of this collaborative testing experience?

Interesting background information

Analyses of group exams, where students take tests or quizzes individually and then in a small group, have been undertaken in a variety of different disciplines (e.g., physiology, nursing, biology, geoscience, and English) and using a range of different exam formats. In virtually every study, exam scores improved, particularly for students in middle- and low-achievement cohorts. Student response to the experience is generally positive, with a sizeable majority reporting that it decreases exam anxiety and promotes deep understanding of the course content.

The study cohort

The cohort included 168 health science students taking a sport and exercise psychology course at an Australian university. The students were at various levels of their undergraduate coursework.

Methodology (a succinct summary)

The collaborative exam experience occurred on the second of three course exams. Students had 45 minutes to complete a 30-question multiple-choice exam. Five of those questions involved advanced knowledge and required higher-order application and analysis of the content. The researchers collected the exams, and students then convened in preconstituted groups that had 30 minutes to do the same exam. The group did not have to agree on an answer. Each student in the group answered the exam questions for themselves. The first exam counted for 80 percent of their grade; the exam completed during group collaboration counted for 20 percent of their grade. Following the exam, students completed an online survey asking about their experiences.

Key findings

  • As a cohort, students’ scores on the exams completed in groups were higher than the exams they did on their own. But high-performing students (those with scores above 90 percent on their individual exams) did not improve as a cohort on their group exams. Those with individual scores below 90 percent in a middle and lower cohort category did improve their scores at statistically significant levels.
  • The whole cohort performed worse on the higher-order thinking questions than they did on the other questions on their individual tests, proving that those questions “were more challenging than the other test questions” (p. 31).
  • The whole cohort improved their higher-order question scores on the collaborative exam, and that included those in the high-performing group: “Students benefitted equally in the collaborative test when answering HOTQ [higher-order test questions], regardless of baseline performance” (p. 31).
  • Most of these students (66 percent) hoped for other collaborative testing opportunities. A majority (67 percent) also indicated the test enhanced their learning (they responded to the survey before they knew their exam scores), and 83 percent felt it made the exam experience more enjoyable.

Cautions and caveats

Research to date, including this study, has not examined the interactions that occur within a group during group collaboration. The assumption is that students are sharing, debating, disagreeing, questioning, and explaining and that these interactions deepen student understanding. It sounds like a solid assumption, but it merits testing. Furthermore, the study is an isolated example: it involves one course at one institution and assesses one collaborative testing model.

Practical implications (what you might want to do about the findings)

First, there’s enough evidence supporting the benefits of collaborative testing to warrant trying it—if not on an exam then on a quiz. Second, a structure like this still means the grade earned is an individual grade and not a group grade. A student who thinks the group is wrong is free to select the answer they believe is correct. Finally, consistent positive response from students demonstrates the effectiveness of collaboration at changing the exam context. In groups, students feel less anxious and report that their understanding of the content deepens. In other words, collaboration increases the learning potential of exam experiences, which makes them an option worth considering.