The University of Southern California (USC) will build on prior work to test a robust model for assessing player content engagement and social interactions within an augmented reality game (ARG). In partnership with No Mimes Media, USC will use machine learning algorithms to make automated player inferences to customize game play. The content focus of the game will span a range of STEM disciplines, with a special emphasis on earth science content and scientific investigation & experimentation reasoning. High school youth from underserved communities in Los Angeles will be recruited to participate in the endeavor.
This pathways project will use various "rabbit hole" techniques to attract freshmen and sophomore students from partner charter schools to the online game. The rabbit hole strategies may include cryptic posters, inquisitive signs, & SQR codes strategically placed in plain and open view of the target group. The game will be fully accessible to the target group online. During the ARG experience, youth players will encounter STEM concepts and scientific problems. Antagonistic characters will promulgate scientific misconceptions and nonscientific reasoning and challenge players to employ their scientific knowledge and skills to level-up, gain badges, and move through the game. As game play persists, machine learning algorithms will gather data on the players learning competencies and social interactions within the game. These data will be aggregated and analyzed to assess learning and interactions within the ARG environment. Additional analyses will be conducted by the mixed methods approach the external evaluation group, CRESST, will employ for the project formative and summative evaluations.
Approximately 300 youth, within the target grade levels, are expected to participate in the gaming experience. However, given that access to the game and assessment tools will expand beyond the target group, the potential reach of the project could be much greater. Further, the stated aim of the project is not only to produce a scalable model for broad implementation but it also endeavors to provide puppetmasters with research and assessment tools to create more individualized experiences and improved learning outcomes for players within ARG environments.
