Can robot-supported learning enhance computational thinking?—A meta-analysis

Abstract

Computational thinking (CT) is crucial for students. Robot-supported learning has emerged as a popular approach for CT cultivation. To understand the effects of robot-supported CT cultivation, the current study conducted a meta-analysis to analyze studies from Web of Science, Google Scholar, and Science Direct databases (2012–2022) which screened using keywords ``computational thinking'' and ``robot''. After further screening, 26 peer-reviewed articles were selected. We synthesized the 33 effect sizes to assess the overall effectiveness of robot-supported learning on CT and its sub-dimensions: concepts, practices, and perspectives. Our findings revealed a medium effect of robot-supported learning on students' CT (g = 0.643), a large effect on CT concepts (g = 0.650), a medium effect on CT practices (g = 0.587), and a huge effect on CT perspectives (g = 1.419). We also explored the moderators (i.e., grade level, study duration, culture, learning strategy, and assessment tool) that might influence the effects of robot-supported CT cultivation. Moderator analyses indicated that short-term (less than four weeks) robot-supported learning had a larger effect (g = 0.901) on CT practice than long-term (more than eleven weeks) learning (g = 0.309). This meta-analysis indicates that educators could apply robot-supported learning to cultivate students' CT and concludes with related implications.