Innovation Suppression and Clique Evolution in Peer-Review-Based, Competitive Research Funding Systems: An Agent-Based Model

Peer review is ubiquitous in modern science: from the evaluation of publications to the distribution of funding. While there is a long tradition of, and many arguments for, peer review as a beneficial and necessary component of scientific processes, the exponential growth of the research community, the 'publish or perish' pressures and increasing insecurity and competition for research grants have led to an increasing number of voices describing the weaknesses of the system. One of the most frequent accusations against the peer review system is that it inhibits true innovation. The availability of better data mining tools allows interested stakeholders, in principle, to monitor many aspects of the process and to promote a better understanding of the interplay of various factors. 'In principle' - because a lot of information is hidden behind the screens of anonymity and confidentiality. Our work presents an attempt at a theoretical understanding of some aspects of the process via an idealized agent-based model, which describes the effects of the peer review done by 'imperfect' agents, in particular with respect to promotion of mediocrity and to formation of self-serving cliques. The results of the model suggest that both phenomena can be quite robust and require careful monitoring of the system to combat their negative effects. Some mitigating measures are simulated and discussed.