The Effect of Programming Teaching on Computational Thinking Skill: A Meta-analysis Study

Abstract

Aim: The present paper seeks the overall effect of programming teaching on computational thinking skill. Method: The present study used meta-analysis method that intends to demonstrate the whole picture by combining the results of experimental studies examining the effect of programming teaching on computational thinking skill. Treatment effectiveness meta-analysis, a type of group comparison meta-analysis, was used in the present meta-analysis study. Findings: Findings regarding sub-problems show that the overall effect of programming teaching is a large and positive effect on students' computational thinking skill (g=0.708, %95 CI [0.555-0.861], p<0.000). Although the applied programming was determined not to be an effective moderator in meaningfully differentiating the effect sizes, of the four programming types whose effects were examined on the computational thinking skill of learners, text-based programming was more effective than the other three programming types. The general effect sizes of studies on computational thinking skill by the years in which the experimental application was carried out demonstrate that the effect size increases over time and that the studies in 2024 have the largest effect size with a very wide effect level. Type of publication moderator reveals that doctoral dissertations has a small effect size and master's theses a large effect size. Educational level moderator presents a huge effect size at pre-school, a very large effect at primary school, a medium effect size at secondary and high school, and a low effect at higher education. Based on these findings, it can be argued that as the educational level increases, the effect of programming teaching on computational thinking skill decreases. However, it can be stated that educational level is an important moderator in differentiating effect sizes. In the study, the sample group with the smallest effect size was the sample group with 26-50 participants with a medium effect, while the highest effect size was the sample group with 0-25 participants with a large effect. The experimental application duration in the independent studies in the present meta-analysis is between three weeks and 18 weeks. It is determined that the most experimental applications (k = 25) are carried out in the 7-10 week period, and the studies carried out between 3-6 weeks have the largest effect size with a large effect size level. It shows that the duration of experimental application is not effective in significantly differentiating the effect sizes of programming teaching on computational thinking skill. Originality: This study is original in that it as a whole examines the effect of studies examining the direct effects of programming types on computational thinking skill.