Abstract
AbstractIn the last two decades, the importance of research on block-based programming education has grown. The use of block-based programming tools is receiving attention not only in computer science courses, but also in robotics education. The effects of such programming on children’s learning outcomes have been examined, but the results have been inconclusive. The purpose of this meta-analysis was to examine the mean effect of block-based programming compared with traditional instruction (i.e., text-based programming) on children’s learning outcomes, including problem-solving skills, programming skills, computational thinking and motivation. The effect size and effects of moderators (publication year, sample size, publication sources and study region) were also examined. The database search yielded 19 publications with 31 effect sizes (n = 1369). Block-based programming had a significantly larger effect size than did traditional instruction for overall learning outcomes. More specifically, we found a large effect size for problem-solving skills, small effect sizes for programming skills and computational thinking, and a trivial effect size for motivation. No moderating effect was detected. Effect sizes for outcomes were large in research conducted in the Americas and Asia, medium in studies conducted in Europe and trivial in studies conducted in the Middle East. No evidence of publication bias in the studies was detected. These study findings support the benefits of block-based programming education for children’s learning outcomes, especially their problem-solving skills. Future research should examine additional dependent variables.KeywordsBlock-based programmingMeta-analysisVisualized programming language
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have