This study operationally defined and measured collaboration and compared the products and structure of collaborative groups that used computer-mediated communication. Key characteristics of collaboration selected from the literature were interdependence, synthesis, and independence, and a model for evaluating these characteristics was developed. All communication in this study occurred via asynchronous computer-mediated communication, using a threaded Web discussion. Participants in the study were graduate students, studying the same course with the same instructor at two venues. The students were divided into small groups from one or both venues, and four of these groups were studied. All students were given a problem to solve involving the cost-benefit trade-offs of distance education. The groups received different instructions. Two of them were told to collaborate on a solution, and the other two were told to select a role and discuss the problem from that point of view. Groups that were instructed to collaborate were more collaborative, but they produced a solution of a lower quality than the other groups. No conclusions could be drawn from the results on the structure of the groups. The role of collaboration in problem solving is discussed along with methods for creating more effective collaboration.