Do editors have a silver bullet? an agent-based model of peer review

This paper presents an agent-based model of peer review that looks at the effect of different editorial policies of referee selection. We tested four author/referee matching scenarios as follows: random selection of referees, selection of referees with a similar status to submission authors, selection of higher-skilled and lower skilled referees. We tested these scenarios against three types of referee behaviour, ie., fair, unreliable and strategic and measured their implications for the quality and efficiency of the process. Results show that in case of randomness of referee judgment, any editorial policy is detrimental for peer review. If referees behave strategically, certain matching policies, such as selecting referees of good quality, might counteract possible bias.