Critical thinking (CT) is among the core abilities students need to acquire to overcome the unprecedented and ever-changing challenges of the 21st century and beyond. To facilitate the attainment of CT, numerous researchers worldwide have conducted empirical studies focusing on leveraging computer-supported collaborative learning (CSCL) to improve students' CT. Notwithstanding the existence of empirical studies, meta-analyses that clearly show the overall effect sizes (ESs) of the impacts are wanting. The current meta-analysis aimed at filling this lacuna by examining 30 empirical studies that were conducted between the years 2014 and 2023, stored in databases such as Scopus, the Web of Science, ERIC and Google Scholar, and search engines such as Google and Ecosia and that met the inclusion criteria. The study included journal articles, conference papers and student theses. The total number of participants in the studies was 2,980, representing 11 countries. The results indicated that CSCL was an effective instructional approach to foster students' CT with statistically significant levels of overall ES estimates (ES = 0.859, z = 8.55, p < 0.001, 95% CI [0.661-1.055]). The results also indicated that there was no apparent evidence of publication bias (Egger's regression = 0.478, p > 0.05). Instructional activities, technology usage, participants' level of education, dimensions of CT, intervention duration and sample size were found to significantly moderate the overall ES estimate. Based on the results, recommendations are made regarding effective ways to use digital technology and collaborative learning for honing students' CT.