March 1, 2013
The Effects of Collaborative Testing
Although letting students work together on exam questions is still not a common instructional practice, it has been used more than might be expected and in a variety of ways. Sometimes students work together in groups; other times with a partner. Sometimes those groups are assembled by the instructor and sometimes students are allowed to select their partners or group members. Sometimes the groups share multiple exam experiences; other times they work collaboratively only once. Sometimes the group submits one exam with everyone in the group receiving that grade; other times students may talk about exam questions and answers but submit exams individually.
Why let students work on exams collaboratively? There are a number of reasons, most of which have been explored empirically. When students discuss questions and possible answers, they intensely engage with the content, which increases the learning potential of an exam experience. The activity develops cooperation and communication skills. But the reason most often given is that working with other students decreases exam anxiety, particularly for those students whose levels of anxiety compromise their ability to perform on exams.
When asked why they don’t use collaborative testing, most faculty report being afraid that students who have not prepared for the exam may inappropriately benefit from the knowledge of students who have studied. Grades should be measures of individual learning.
Based on a review of the literature on collaborative tests, these authors decided to explore four questions in their study: 1) How does collaborative testing affect test scores? 2) Is anxiety related to the effectiveness of collaborative testing? 3) Which students benefit from collaborative testing? and 4) How does the quality of interaction within the groups affect test performance? (p. 165)
The study included some unique design features. To deal with the potential problem of students coming to the exam unprepared and thinking the group would pull them through, students were told to prepare for the exam as if they were taking it individually. Those who would be taking the exam collaboratively would be randomly selected at the beginning of the period. On test days, those selected to take the exam collaboratively were moved to another room. Their group interactions were observed by a proctor who evaluated the quantity of interaction, the level of enthusiasm, and the degree of give-and-take displayed by the participants. (p. 166) The groups did not have to agree on answers, although they could change their answers based on discussions that occurred within the group. Tests were still submitted and graded individually.
As for how the collaboration affected test scores, the results were positive. “Collaborative testing was more successful for a significant majority of students than was traditional, individual testing, although the advantage (3.83%) was smaller than found in some previous studies.” (p. 172) Based on standardized measures of test anxiety, “students with higher initial test anxiety scores were most likely to benefit from the collaborative testing procedure and to show a decrease in test anxiety when taking tests collaboratively.” (p. 172) And finally, “high interaction scores, both proctor and student-rated, were related to better performance under the collaborative condition.” (p. 172)
Consistent in the research on collaborative testing mechanisms is the favorable response they generate from students. Students regularly report enjoying the experience and indicate they would choose it in the future, if given the opportunity. In this study, students said the testing mechanism was beneficial because it helped them develop good working relationships with classmates and helped them remember material they had forgotten. There were students in this study who did perform better on the tests they took individually, and a third of those students thought they did better on the individual exams because they found the discussions with other students confusing.
The researchers conclude that collaborative testing is a “worthwhile technique.” “For, even if overall test performance is not greatly improved by collaborative testing, the positive experiences of increased camaraderie and anxiety reduction it engenders could constitute substantial long-term benefits in the form of increased confidence, motivation, and willingness to continue one’s education.” (p. 173)
Reference: Pandey, C., and Kapitanoff, S. (2011). The influence of anxiety and quality of interaction on collaborative text performance. Active Learning in Higher Education, 12 (3), 163-174.
Reprinted from The Effects of Collaborative Testing, The Teaching Professor, 26.1 (2012): 3, 4.