Data Analytics of Mobile Serious Games: Applying Bayesian Data Analysis Methods

The mobile serious game Held functions as a highly effective pedagogical tool for teaching resuscitation theory and practical skills to secondary school students. By integrating game-based learning with traditional instruction, the simulation achieves knowledge transfer outcomes comparable to formal classroom training. The research confirms that these digital interventions remain effective regardless of a participant’s prior gaming experience, establishing mobile simulations as a viable, scalable alternative for essential medical training in educational settings.

Methodologically, the study employs empirical Bayesian data analysis, utilizing Beta distribution models to navigate the complexities of uncontrolled variables such as student motivation. This analytical framework proves superior to traditional frequentist approaches, particularly in its ability to isolate performance metrics from subjective human evaluation. By synthesizing diverse data sources—including manikin sensor readings, longitudinal retention surveys, and game-log metrics—the analysis provides a robust, nuanced view of student progress that traditional testing methods often overlook.

A critical finding of this research is the identification of a systematic, unexplained bias against female participants within expert observer ratings. Despite this external bias, objective performance data reveals that female students consistently outperform their male counterparts within the game environment. Furthermore, the study suggests that visual design choices, specifically the use of cartoon-like aesthetics, significantly enhance instructional focus and learning outcomes. Ultimately, the application of Bayesian methods not only validates the efficacy of serious games but also serves as a diagnostic tool for uncovering inequities in traditional assessment practices, providing a clear path forward for the design and evaluation of future educational technologies.