Latent Class Analysis of Gameplay Metrics from Youth Playing a Robotics Game

Latent Class Analysis of telemetry data from the educational robotics game Robot Champions reveals critical insights into youth engagement patterns and the efficacy of game design. By analyzing 59 days of activity from over 150,000 users, the research identifies four distinct player personas: Fully Engaged, Engaged in Training, Engaged in Building, and Engaged in Driving. These classifications provide a high-resolution alternative to subjective self-reporting, allowing for a more objective assessment of how players interact with complex engineering-based mechanics.

The data highlights a significant challenge regarding player retention, as the vast majority of the user base engages for only a single session lasting less than four minutes. This trend suggests a steep learning curve and a lack of sustained interest in competitive features. Furthermore, the behavioral patterns indicate that players frequently bypass traditional tutorials in favor of immediate hands-on experimentation. This preference for active exploration over structured instruction points to a need for more intuitive in-game scaffolding, contextual hints, and refined reward structures to mitigate early-stage frustration.

Ultimately, these findings demonstrate that objective gameplay metrics serve as a powerful diagnostic tool for developers aiming to improve game design and support STEM skill development. By aligning game mechanics with the observed behavioral realities of the audience, developers can better facilitate persistence and deeper learning. Future efforts should focus on integrating more granular telemetry data to further refine these behavioral models, ensuring that educational games effectively balance technical complexity with user accessibility to foster long-term engagement.