摘要：With technological advances, it is now possible to use games to capture information‑rich behaviours that reveal processes by which players interact and solve problems. Recent problem‑based games have been designed to assess and record detailed interactions between the problem solver and the game environment, and thereby capture salient solution processes in an unobtrusive way (Zoanetti, 2010; Bennett et al., 2003; Shute & Wang, 2009). The Assessment and Teaching of 21st Century Skills project used an innovative approach to capture these processes through responses (activities and communication patterns) within collaborative games (Griffin & Care, 2015). Game player response patterns are identified as behaviours that are indicative of skills and are captured in real‑time game play within time‑structured log files. The analysis of these log files allows for inferences to be drawn in regard to the efficiency and quality of player performance. A concern with this approach is that game development for this purpose is complex, time consuming and expensive, with unique scoring applied to each game. This paper presents another approach that identifies, across games, common behaviours. A systematic scoring system for assessing player behaviours in games could provide access to useful data not only from new games but from existing games. This paper presents such an approach using collaborative games situated in a problem‑solving context.
关键词：Collaboration ; problem solving ; online assessment ; log stream data ; measurement ; e-learning
