學習電腦程式對臺灣學生運算思維能力影響之後設分析

Meta-Analysis of the Effect of Programming Instruction on Students’ Computational Thinking in Taiwan

「運算思維」（Computational Thinking, CT）一詞為J. Wing所提出。自CT概念被提出後，有些學者也指出學習程式語言可以提升學生的認知能力（如創造力、理解力、批判思考能力、問題解決能力等）。本研究之主要目的為探討學習程式設計的經驗是否能提升臺灣學生之CT能力，並進一步檢視有哪些因素可能會調節整體的成效。本研究經搜尋國際及臺灣各資料庫，最後共有28篇研究（45組ES）符合標準。結果顯示，固定效果模式（fixed effects model）與隨機效果模式（random effects model）之整體平均ES（ES+）各為0.46（z = 13.96, 95% CI = 0.40-0.53, p < .001）與0.54（z = 7.59, 95% CI = 0.40-0.68, p < .001），可見無論是採用哪一種模式，都顯示程式設計教學對學生CT的成效顯著優於無程式設計的教學。其次，在調節變項分析方面，八個調節變項中有三個的QBetween（QB）達顯著差異：一、發表形式，期刊論文顯著高於學位論文；二、程式語言，Scratch、Micro:bit以及混合組都顯著優於Code.org；三、融入策略，融入協同學習的研究成效最佳，且顯著地優於機器人輔助、其他與無策略。本研究的結果提供各級教師一個以累積研究為本的有效提升學生CT的教學方式。

 

Computational thinking (CT) was first introduced by J. Wing. Since then, several researchers have suggested that learning programming languages can enhance students’ cognitive abilities, including their creativity, comprehension, critical thinking, and problem-solving. The current study investigated whether learning programming can enhance Taiwanese students’ CT abilities and explored factors that may moderate this effect. International and Taiwanese databases were searched for relevant articles, and 28 studies (45 effect sizes) met the predetermined inclusion criteria. The overall mean effect sizes for fixed- and random-effects models were 0.46 (z = 13.96; 95% confidence interval: 0.40-0.53; p < .001) and 0.54 (z = 7.59; 95% confidence interval: 0.40-0.68; p < .001), respectively. Thus, in both models, programming instruction significantly outperformed nonprogramming instruction in improving students’ CT performance. In a moderator analysis, three of eight moderators exhibited significant Qbetweens (QBs): (1) type of publication– journal articles exerted stronger effects than did theses; (2) programming language– Scratch, Micro:bit, and mixed groups all significantly outperformed Code.org; and (3) instructional strategy– studies integrating collaborative learning achieved the best outcomes, significantly outperforming those using robot-assisted, other, or no strategies. This study provides educators at all levels with evidence-based guidance for effectively enhancing students’ CT abilities.