Group awareness (GA) is essential for computer-supported collaborative learning (CSCL), as it informs learners about other group members' activities, knowledge, and emotions. A key advantage of GA support is that it can collect, process, and visualize GA information, which provides a basis for students' reflection and adjustment during collaborative learning, thus facilitating their learning performance. However, empirical findings regarding the effectiveness of GA support have been inconsistent. The present study conducted the first three-level meta-analysis of 46 empirical studies to examine the effects of GA support on students' learning performance and further explore the possible moderating factors that may have contributed to the inconsistencies of primary studies. The results indicated the following: (1) GA support in CSCL had a moderate significant effect on students' learning performance (Hedges'g = 0.46, p < 0.001); (2) GA support in CSCL had the greatest influence on students' cognitive development (Hedges'g = 0.49, p < 0.001), followed by behavioral participation (Hedges'g = 0.47, p < 0.001), and then social emotion (Hedges'g = 0.38, p < 0.001); and (3) GA support type and group size were the only two significant moderating factors. Based on these findings, we propose several theoretical and pedagogical implications.