<p>We investigated the effectiveness of the cooperative base groups (CBGs) technique in facilitating cooperative learning in small WhatsApp groups for first-year Computer Science students. A multi-level social network analysis (SNA) and Pearson’s correlation coefficient was embedded within a concurrent embedded mixed-methods framework to answer the research questions. At macro level, none of the whole development networks (WDNs) were complete, and six were fragmented. At meso level, cliques and components were found. At micro level, some of the students expanded their personal development networks (PDNs). A weak positive correlation was found between the size of the PDNs and final marks. The results suggested that those who expanded their PDNs tended to perform better than those who did not. Previous assumptions about network centrality and academic achievement could not be supported, as many of the most central nodes were not the top academic performers. A high positive correlation was found between the size of PDNs and the final marks of students who failed. The CBG technique was effective in facilitating cooperative learning in WA groups, but we recommend frequent SNA to identify at-risk students, longer-term implementation of the technique, and further investigation into the instructor’s role in promoting cooperative learning.</p>