A social network analysis of the citation diffusion of academic articles

This research aims to investigate whether the diffusion of academic articles’ citations follows a random order. A social network analysis (SNA) is employed whereas articles’ citation networks were constructed based on multiple generations of citations. For example, the “first-generation citations” represent articles citing the original publication, whereas the second generation represents those citing the first-generation’s citations. Results revealed a heterogeneous hub-and-spoke network with relatively low density. In this network only a few influential articles within small communities appear to control the information diffusion through the network, lending strong support to the “Matthew Effect”, which is a variation of the “rich get richer” model. Such articles seem to form bridges between key clusters in the network through information brokerage in the network. The “small world preferential attachment” network linking the cited articles was also confirmed through statistical robustness checks.