Simulation and manipulation play vital roles in teaching procedural knowledge in science, technology, engineering, and mathematics (STEM) education. However, most studies in these fields have only focused on learning effectiveness rather than on students' collaborative problem-solving (CPS) processes and cognitive processing patterns in problem-solving activities. This paper aims to analyze learners' cognitive processing patterns in a CPS teaching activity that integrates Facebook discussion tools and simulation-based teaching software. The participants in this study included 45 students from the College of Management at a university in northern Taiwan. These students took an elective course titled "Introduction to the Network." This study used a quantitative content analysis (QCA) to code the online discussions and to explore the represented cognitive levels. Next, a lag sequential analysis (LSA) was conducted to analyze the coding results of the QCA to assess the behavioral patterns during the discussion process. The results indicate that students' network troubleshooting abilities improved over the course of the study. The learners' discussion patterns and cognitive levels during the activity are analyzed and discussed here. Differences in sequential cognitive processing patterns were found between the "discussion-centered" and "manipulation-centered" groups. The findings of this research may serve as a reference for instructors of STEM classes when integrating social networking services (e.g., Facebook) and CPS when designing virtual learning activities to improve learning effectiveness and increase cognitive levels. (C) 2014 Elsevier Inc. All rights reserved.
