Using Log and Discourse Analysis to Improve Understanding of Collaborative Programming

Abstract

The importance and ubiquity of computing and computational thinking (CT) is leading to re-design of K-12 curricula and the development of appropriate platforms to support computer science instruction. One such environment is the block-based, synchronous programming environment, NetsBlox, that allows for real-time collaboration amongst students. This work presents a novel combination of discourse and activity log analyses to study the collaborative behaviors of K-12 students as they worked on a week-long cybersecurity curriculum. Groups of students were assessed based on pre-post-test learning gains in cybersecurity and CT. We analyze the differences between the collaborative behaviors and discourse of high and low performing groups using case study and differential sequence mining analyses to characterize productive and unproductive collaborative problem solving in programming tasks.