Collaborative Group Versus Individual Assessment in Mathematics: Group Processes and Outcomes

Several states, such as Connecticut and California, are attempting to incorporate group assessment into their large-scale testing programs. One intention of such efforts is to use scores from group assessments as indicators of individual performance. However, a key technical question for such assessments is “to what extent do scores on a group assessment actually represent individual performance or knowledge.” This study by UCLA professor and CRESST researcher Noreen Webb sheds some light on this substantial technical question. Webb gave two seventh-grade classes an initial mathematics test as a group assessment, where exchange of information and assistance was common. Several weeks later, she administered a nearly identical individual test to the same students where assistance was not permitted. The results showed that some students’ performance dropped significantly from the group assessment to the individual test. These students apparently depended on the resources of the group in order to get correct answers and when the same resources were not available during the individual test, many of the students were not able to solve the problem. Webb concluded: “Scores from a group assessment may not be valid indicators of some students’ individual competence. Furthermore, achievement scores from group assessment contexts provide little information about group functioning.” Webb’s study suggests that states or school districts who intend to assign individual scores based on group assessments may want to seriously rethink their intentions.