The Effects of Artificial Intelligence Education: A Meta-Analysis

Abstract

Purpose: According to the government policy, interest in and research on artificial intelligenceeducation has been actively conducted in Korea since 2019. The purpose of this study was to synthesizeeffects of artificial intelligence education through meta-analysis. Methods: The researcher rigorously selected a total of 43 primary studies including 25 journal articlesand 18 theses and dissertations reporting effects of artificial intelligence education. To calculate theeffect size, first, the effect size for each study was calculated, and then the overall effect size wascalculated. Subgroup analyses were then performed to compare differences by study characteristics. CMA 3.0 program was used in data analysis. Results: First, the overall effect size of artificial intelligence education is 0.687, which can beinterpreted as a medium effect size. Second, effect sizes according to the school level were 0.947for college, 0.784 for high school, 0.784 for middle school, and 0.633 for elementary school. Third,effect sizes according to the learning environment were 0.915 for online tools, 0.777 for the unplugged,0.705 for physical computing, 0.684 for block coding, 0.493 for script coding, and 0.332 for others. Fourth, effect sizes according to dependent variables were 0.857 for the cognitive domain and 0.57for the affective domain. Conclusion: Based on the results of this study, it is expected that school administrators and educatorswill be able to develop effective and efficient artificial intelligence education programs. Therefore,the author hopes that artificial intelligence education will enable students at all school levels toexperience successful learning and achieve their educational goals.