Measuring Collaborative Behavior in COMPS Student Problem-Solving Discussions

This project attempts to observe and measure collaboration between students who are working together to solve problems in a computer programming class. Collaborative problem-solving is among the skills that K-12 education is endeavoring to teach and measure, and appears in the most recent PISA international comparison of educational achievement. In COMPS (Computer Mediated Problem Solving) exercises students work together via online typed-chat. The communication is recorded and stored for further analysis. Using student dialogue files, the main research activity consists of manually classifying student dialogue according to four categories of collaborative utterance. The four categories are: sharing ideas, negotiating ideas, regulating problem solving, and maintaining communication. In concert with other researchers, we are using the hand-labeled data to attempt to train computer text classifiers to identify these behaviors. From there we will count the different behaviors and look for patterns of interaction. This is expected to reveal the conversational fingerprints which are characteristic of successful and unsuccessful student collaborations. This research may be useful to more easily measure collaboration within a group and to obtain a better understanding of how collaboration aids understanding and learning