Enhancing AI Education Through Educational Games: Evidence From a Meta-Analysis Study

The purpose of this paper is to examine the effects of educational games in AI education through a meta-analysis of 22 empirical studies published over the last decade. We examined the effects of different school segment, experimental period, experimental scales, game types, knowledge types, and teacher interventions on students' knowledge achievement, motivation, and thinking skills. The results of the study proved that educational games have an overall positive impact on AI education. Moderator analyses indicated that the games have a greater impact on students' motivation and knowledge achievement, and a lesser impact on thinking skills. Unexpectedly, the longer the experimental period, the smaller the experimental period, and the program with teacher intervention, the better the teaching and learning outcomes. In terms of game types, puzzle and robotics games were the most effective. And on the type of AI knowledge, the study found that games were most effective when used as content to teach AI algorithms.