Game over? Investigating students’ working memory, situational interest, and behavioral patterns as predictors of dropout in an educational game

Abstract

Educational games are designed to increase students' motivation and persistence. Nevertheless, student dropout is a very common issue that many game interventions in education face, which can hamper students' learning opportunities. Dropout can be manifested in different ways for different reasons, for instance quitting due to lack of interest or failing due to lack of abilities to succeed in the game. Thus far, little is known about the underlying factors that lead to student dropout. The present paper (N = 272 early secondary school students) investigates different types of dropout and whether students' working memory capacity and situational interest are predictive of the likelihood of dropping out during an educational game using augmented reality (AR). Moreover, log data were used to explore differences in students' in-game behavior (i.e., number, type, and patterns of mistakes) while playing the game. The results indicate that 16.9% of the students dropped out during the game, either because they quit or because they failed to reach the last level. Working memory capacity and situational interest were not predictive of students' dropout. However, process maps of students’ in-game behavior showed that dropout students differed in the number of mistakes they made and also followed a different behavioral pattern than non-dropout students. In all, this study contributes to the knowledge base by revealing different types of dropout students who showed distinct behavioral patterns. More insights into these patterns could help to identify students at-risk of dropping out early on and to develop targeted interventions to prevent dropout.