This study investigates the collaboration networks within the social work discipline using the recently established Social Work Research Database. By means of network-analysis methods, we investigate the collaborations between 1990 and 2014, while the nodes of this network represent social work scholars (N = 19,789) connected by edges representing common publications (N = 314,180). We analyse the structural characteristics of these co-authorship networks and how they changed over time. The results show an increase in scientific collaborations among social work scholars and an evolution from mainly disconnected groups to a collaborating scientific community, while the network shows a small-world structure. The findings also reveal the existence of the so-called Matthew effect within the social work discipline. Specifically, nodes with a high number of connections are associated with a higher probability of acquiring new connections than other nodes, as well as becoming hubs. We therefore argue that an inequality is constantly reproduced and growing over time that implies barriers especially for early-career researchers.