<p>Unplugged activities (UA) and plugged activities (PA) are two primary teaching approaches used to develop students' computational thinking (CT). However, reaching an academic consensus regarding which approach is more effective remains elusive. This study presents a meta-analysis of 37 studies published between January 2006 and March 2025. These studies are used to compare the effectiveness of UA and PA for developing students' CT. The results indicate that, overall, PA (g = 0.606) is more effective than UA (g = 0.501) in developing students' CT. This study also compares the effects of UA and PA across the four moderator variables of educational level, intervention duration, type of course, and learning model. The results show that UA is better than PA in terms of developing students’ CT in game-based learning. However, PA is better than UA in secondary school and college, in interventions lasting less than four weeks, humanities and arts courses, and problem-based and project-based learning. The effect of UA and PA on students' CT development may be quite similar in the primary education stage, when the intervention duration exceeds four weeks, in information and technology courses, and in natural science courses.</p>