Phase Transitions in Peer Review Systems Arising from Strategic Interactions

Peer review plays a central role in academic publishing and grant evaluation, serving as a mechanism for quality control and resource allocation. In this process, individuals alternate between reviewer and submitter roles, making binary accept–reject decisions on others' work. Such role-switching naturally gives rise to repeated strategic interactions, where decisions depend not only on paper (or proposal) quality but also on prior community behavior. A common strategy resembles "tit-for-tat" in game theory, where individuals begin cooperatively and then mirror others' actions. Yet, the collective dynamics emerging from these interactions remain poorly understood. We develop a theoretical framework that models peer review as a system of repeated strategic interactions encoded in an Ising-like Hamiltonian, incorporating paper quality, personal bias, and reciprocity. The model reveals a phase transition under blind review: small changes in strategic parameters can abruptly shift the community from a cooperative, high-quality regime to an unstable, low-quality one. In contrast, open review removes this transition, producing smoother and more stable dynamics. Comparison with real-world journal data suggests that most communities operate in a hybrid regime, balancing reciprocity with quality-based evaluation. Our findings highlight a trade-off: blind review can promote innovation but risks instability, whereas open review fosters a more equitable and resilient scientific ecosystem.