Building Automated Detectors of Gameplay Strategies to Measure Implicit Science Learning

Educational games have the potential to be innovative methods of assessing learning. This research combines video analysis and educational data mining to measure the implicit science learning that takes place in games. By studying the video data from high school learners playtesting the game Impulse, we observed strategic moves that are consistent with an implicit understanding of relevant science concepts and reliably coded those moves in a sample of 69 high school students. This paper reports on work in progress to use educational data mining analyses that leverage coded video segments to build automated detectors of strategic moves from game log data.