Using Data Science to evaluate Game-Based Learning in informal contexts

The emergence of digital Game-Based Learning (GBL) has sparked interest inassessing its efficacy. This assessment needs to consider the complex mix of narrative andinteractivity typical of video games, which makes it difficult to evaluate to what extent avideo game achieves its stated learning objectives. This challenge is exponentially increasedwhen gaming sessions happen spontaneously in informal contexts, without any supervisionby educators or the option to assess the players' prior knowledge and skills. This workpresents a methodology for analyzing GBL experiences based on data science and the datacollection functionalities offered by current game development platforms. This strategy isapplied to the analysis of a social media simulator designed to promote information literacywithin the video game Julia: A Science Journey. The system collected data on 436 sessionsfrom 112 unique players over six months. The records included information on replayability,identification of fake news, and reaction times. The results suggest that players becomemore adept and swifter at identifying fake news through repeated games. Success inidentifying misinformation is also related to the topic, with hoaxes related to scientificcontent being more easily recognized than those associated with political controversies.