In my role as editor, I have the ability to rescind a review that is 'inappropriate', and the definition of inappropriate is left to my discretion. I have never used this option, and have chosen to deal with the occasional inappropriate review comments in other ways.
For example, in a recent case, an otherwise acceptable review in terms of content and tone had a completely inappropriate statement that demonstrated the discomfort of the reviewer, who was from a western European country, with the government of the author's native country, which is east of western Europe. The review comment had absolutely nothing to do with the scientific content of the manuscript, and the reviewer's recommendation was bizarre. I sent the review to the manuscript authors, but mentioned in my editor's comments that the authors could disregard that particular recommendation by the reviewer.
I recently heard about a more egregious case than anything I have encountered. If I ever received such a review as an editor, I would rescind it. In this case, a reviewer, angry that the manuscript under review disagreed with some of his prior work on the topic, recommended that the manuscript be rejected and included an explicit racist comment about the first author. The editor rejected the manuscript on the basis of this review.
I'm not involved in the research and have no direct interest in the research, but I am familiar with the work in the rejected manuscript and know it to be of high quality. It is shocking that the manuscript was rejected on the basis of a review that clearly demonstrated an extreme lack of objectivity. Just as shocking is that an editor would let an explicitly racist review be sent to the authors.
This is not about political correctness. This is about being ethical. This is about being professional. This is not a marginal case in which a reviewer unfortunately chose to express himself in an awkward way. This is a disgusting case of a reviewer who was so lacking in objectivity about a perceived assault on his exalted reputation that he responded by denigrating an author's ethnicity. And he somehow felt comfortable doing so. The editor's behavior confirms that the reviewer was justified in this feeling of comfort.
I think that the peer-review system works quite well overall, but there are instances that seem to argue otherwise.
Reviewers and manuscript authors don't have many 'rights'. Authors would like to have the right to a rapid and fair review. Reviewers would like to have the right to have their comments taken seriously by authors and editors. In real life, the review process can take days to years; reviewers may or may not be as objective as they could be; authors may or may not take seriously the comments of reviewers when revising manuscripts; and editors may make decisions by flipping a coin or consulting their pet beagle.
I think that reviewers who include explicitly offensive comments about an author's race or other personal characteristics (gender, religion etc.) lose the 'right' to have their review sent to the author. This seems reasonable to me, and doesn't infringe at all on the time-honored tradition of implicitly or explicitly insulting the intelligence of the author.
7 months ago