On Collective Dynamics of Academic Journals: Polarization, Fluctuation, and Robustness

The dynamics of academic journals are observable from time series of citation indices, such as journal impact factor (JIF), which are quantitative emergence of the citation network topology ever growing. It is meaningful to study the variation of citation indices and quality of journals through computational studies. In this article, an experimental system is developed to generate virtual citation networks by simulating key social activities, including manuscript submission, peer review, journal publication, and article citation. The system is built upon a model grounded solely in basic individual behavioral dynamics and free from reliance on specific technical assumptions, ensuring broad generality. By incorporating bibliometric attributes such as article quality scores and temporal publication data, the system enables replication of journal behaviors and citation dynamics. A series of computational experiments are conducted on the system, revealing and analyzing various phenomena under different scenarios. Key findings include the identification of dynamic patterns such as stationarity, inertia, and polarization tendencies in journal performance over time. The experiments further demonstrate the robustness of journal dynamics against moderate external disturbances, such as self-citation behavior, and highlight a gradual weakening of volatility in journal citation indices. These findings are further explained through a semianalytical approach. The study offers actionable insights for journal management, policy design, and long-term evaluation strategies, contributing to a deeper understanding of the temporal evolution of academic journal communities.