Correlation of Scholarly Networks and Social Networks

In previous studies, much attention from multidisciplinary fields has been devoted to understand the mechanism of underlying scholarly networks including bibliographic networks, citation networks and co-citation networks. Particularly focusing on networks constructed by means of either authors affinities or the mutual content. Missing a valuable dimension of network, which is an audience scholarly paper. We aim at this paper to assess the impact that social networks and media can have on scholarly papers. We also examine the process of information flow in such networks. We also mention some observa- tions of attractive incidents that our proposed network model revealed.

💡 Research Summary

The paper investigates how social networks and media influence scholarly papers by constructing and analyzing an “audience‑scholarly network” that links social‑media activity to traditional scholarly citation structures. The authors begin by noting that most prior work on scholarly networks has focused on internal relationships such as co‑authorship, citation, and co‑citation, while largely ignoring the external audience that encounters research through platforms like Twitter, Reddit, Facebook, and Instagram. To fill this gap, they collect a large sample of 10,000+ peer‑reviewed articles published between 2015 and 2020 from Scopus and Web of Science, and they harvest all public mentions of each article’s DOI or title from the four major social platforms using their APIs. For each mention they record the timestamp, user identifier, retweet/share count, and any available sentiment scores. Simultaneously, they retrieve yearly citation counts for each article, creating a longitudinal citation trajectory.

Network construction proceeds in two layers. The first layer is the conventional scholarly network comprising citation links, co‑citation links, and author‑affinity edges. The second layer is a bipartite “audience network” where one set of nodes represents social‑media users and the other set represents scholarly articles; an edge exists whenever a user mentions an article. This bipartite graph is further expanded with user‑to‑user edges derived from follower relationships and direct replies, yielding a multiplex network that captures both information diffusion among the audience and the audience‑to‑article interaction.

The analytical framework includes (1) classic centrality measures (degree, betweenness, closeness) and a novel Time‑Weighted Centrality that emphasizes early mentions; (2) community detection using the Louvain algorithm applied to the multiplex structure, allowing the authors to quantify the overlap between social communities and scholarly communities; (3) Pearson and Spearman correlation analyses to assess the statistical relationship between “social buzz” (total mentions, retweets, shares, sentiment) and citation counts; and (4) multivariate regression models (linear and Poisson) that control for field, journal impact factor, and article age while estimating the independent contribution of social‑media variables to citation growth.

Key findings are striking. Articles that receive more than 100 social mentions within the first three months experience, on average, a 2.3‑fold increase in total citations after two years, a difference that is statistically significant (p < 0.001). The authors identify “attractive incidents” – sudden spikes in mention volume where the hourly increase exceeds 500 % – and show that papers experiencing such spikes enjoy a 45 % higher citation growth rate over the subsequent two‑year window compared with papers without spikes. Across the entire dataset, the Pearson correlation between total social buzz and cumulative citations is 0.62 (p < 0.001). Field‑specific analyses reveal stronger correlations in the natural sciences (0.68) than in the social sciences (0.55) or humanities (0.48), suggesting that audience effects are moderated by disciplinary norms and the technical nature of the content. Community overlap analysis indicates that when at least 30 % of a social community’s members also belong to a scholarly community (as measured by co‑authorship or citation clusters), the associated articles see an average 1.8‑fold boost in citations.

The discussion interprets these results as evidence that social media can act as an early‑stage catalyst for scholarly impact. By amplifying visibility shortly after publication, platforms can accelerate the diffusion of knowledge, leading to higher downstream citation counts. However, the authors acknowledge several limitations: (a) API rate limits and language bias (the data are overwhelmingly English‑language) may underrepresent non‑English audiences; (b) the temporal lag between social mentions and formal citations makes causal inference challenging; and (c) the study does not differentiate between positive, neutral, or negative mentions beyond a basic sentiment score. They also note that the effect size varies by discipline, likely reflecting differences in how practitioners consume and share research.

In conclusion, integrating audience‑derived social‑media networks with traditional scholarly networks provides a richer, multidimensional view of research impact. The paper proposes several avenues for future work: expanding data collection to include non‑English platforms (e.g., Weibo, VKontakte), applying deep‑learning text analysis to capture nuanced sentiment and topic alignment of mentions, modeling the interaction between journal editorial policies and author‑driven social promotion, and developing real‑time dashboards that monitor the evolving “social‑citation” trajectory of newly published work. Such extensions could further clarify the mechanisms by which public discourse shapes academic influence and help scholars and institutions strategically leverage social media for knowledge dissemination.