Crossing the Bridge: Connecting Game-Based Implicit Science Learning to the Classroom Elizabeth Rowe, Jodi Asbell-Clarke, Erin Bardar, Emily Kasman, & Barbara MacEachern Educational Gaming Environments (EdGE) group @ TERC Abstract: Games offer a unique opportunity to promote and study implicit learning that could be foundational for further STEM learning. This paper reports results from a national classroom implementation study. In this study, 42 teachers were assigned to the Bridge (classroom activities and game play), Games (game play only), or Control group. The game, Impulse, immerses players in the physical laws of Newtonian motion. Hierarchical linear modeling of data from the 23 teachers (50 classes) to complete the study shows a significant positive effect of the Bridge group compared to the Control group on student’s post-assessment scores after accounting for pre-assessment scores. This group effect, however, was significantly moderated by whether or not the class was a Honors/AP class. These initial findings support our conjecture that Impulse can help prepare some students for improved science learning in class. Overview This paper explores implicit science learning through the game, Impulse, and how that learning might be leveraged to improve classroom learning. The assumption inherent in the design of Impulse is that by building foundational learning through games, students may be better prepared to study related science in class. We are testing that conjecture through an implementation study with hundreds of high school students. In this paper, we present the results of pre-post assessments of Newton’s laws of motion from three groups of high school students. The Game group (209 students in 21 classes) was encouraged by their teachers to play Impulse outside of class and some played in class. The “Bridge” group (179 students in 18 classes) was encouraged to play the game outside of class, some played the game in class, and their teachers used examples from the game in their classroom instruction on Newtonian mechanics. One hundred eight (108) students in 11 classes in the Control group neither played the game nor had game examples in class. Researchers gathered pre and post assessment results as well as daily logs of teacher activities. HLM analyses look at the relationship between study group—Bridge, Game, Control—and students’ post-assessment scores. Results show higher post-assessment scores for students in the bridge classes, as compared to the control classes after accounting for the students’ pre-assessment scores, student characteristics, teacher characteristics, and school demographics. Whether or not the class was a Honors/AP class, however, moderated the study group effect. This confirms our conjecture that Impulse can help prepare some students for improved science learning in class. Making the Bridge from Games to Classroom Implicit learning may be foundational to all knowledge (Polanyi, 1966) but is particularly challenging to measure because it is, by definition, largely unexpressed by the learner. Polanyi described implicit knowledge as foundational for explicit knowledge building (1966). McCloskey (1983) and diSessa (1993) studied learners’ misconceptions in science in terms of their implicit understandings of the physical world around them, pointing out that many learners have inaccurate fundamental implicit models of forces and motion because of our daily experiences with friction and other complicating factors. Games offer a unique opportunity to promote and study implicit learning. They are a highly engaging environment that can be designed to use mechanics that mirror authentic science and are rich research environments because of the digital log data they generate (Asbell-Clarke et al., 2012; 2013; NRC,2011). The data exhaust from games (Halverson et al., 2012) can be used along with education data mining methods (EDM) to detect patterns of play that learners use and how those patterns change as players advance towards more successful and sophisticated activities (Baker & Yacef, 2009; Martin et al., 2013).