The shadow of reciprocity on confidential vs. open peer review: An agent-based model

This paper presents an agent-based model of peer review that looks at potential con-sequences of reciprocity strategies by scientists on the quality and efficiency of the process. We analyzed scientist strategies in two conditions: (i) when peer review is confidential and anony-mous; (ii) when peer review is open. We built three scenarios through which we manipulated reviewer behavior: reliable (reviewers always cooperate by providing pertinent opinion), unre-liable (reviewers always defect by providing biased opinion) and strategic (reviewers overvalue other’s work when previously published and vice versa) reviewers. Results showed that the effect of reciprocity strategies on the quality of peer review might be detrimental. This is even more dramatic when peer review is open, as scientists could play Tit-for-Tat direct reciprocity against other colleagues, which may dramatically biases their personal judgement.