Game-based assessment: an integrated model for capturing evidence of learning in play

This paper presents the assessment data aggregator for game environments (ADAGE), a click-stream data framework designed to test whether click-stream data can provide reliable evidence of learning. As digitally-based games become increasingly prevalent in both formal and informal learning environments, robust research and assessment vehicles within well-designed games become vital. A central challenge with this video games research in education is to demonstrate evidence of player learning. Rather than ignore the motivating and information-rich features of games in capturing learning, assessment designers need to attend to the ways in which game-play itself can provide a powerful new form of assessment. This requires learning researchers to think of games as both intervention and assessment; and to develop methods for using the internal structures of games as paths for evidence generation to document learning. ADAGE consists of two main layers: 1) the semantic template that determines which click-stream data events could be indicators of learning; 2) the learning telemetry that captures data for analysis. This study highlights how ADAGE was implemented in science game to provide a wealth of in-game player data indicative of both gameplay and learning progress to reveal important relationships between kinds of success, failure and learning in the game.