Peer review lacks indicators, which is incongruent with its importance. This article reports the reliability of mutual, anonymous review of masters research proposals. Twelve masters students in Information Science refereed anonymously their peers' research proposals according to 7 evaluation items, using a Likert-6 scale. Reliability was expressed as intraclass correlation indices between 0.500 and 0.202, with 9 positive and 3 negative values - ​​lower than usual in professional processes. Reliability is not an absolute measure of quality, but calculating reliability allows for a systematic study of the quality of peer review.