I then was bored with Reviewer B’s somewhat serious and earnest recommendations, and thought that giving the guidance of a bad reviewer may illustrate my feelings a little more. So, Reviewer B, good guy. Reviewer C, bad guy. (Just so I’m clear.)
How to write a review, by anonymous reviewer C.
- Your default position should be rejection – start out with a view to rejecting the paper. Then work backwards from that. Anything else would be sloppy unscientific thinking. You should view reviewing as hypothesis testing. Correct for multiple comparisons: more reviews, more rejections.
- Remember, you’re not marking, or grading a piece of coursework. You don’t need to give constructive feedback, nor justify your impression. The author will not knock on your door with a sad face asking for an explanation of the review. On the other hand, there will be articles which have been written sloppily and without care, and then you will find it easiest just to slip into ‘teaching mode’ and complete your review as if reading an undergraduate essay.
- Always start the review with a catty insult. You will find online dictionaries of insults. The best insults are those which amuse yourself. They are those which contain archaic language and speak to a weary, world-wise ennui, and they are invariably sarcastic. ‘I cannot bring myself to accept this paper. Nonetheless, I found it to be novel. It is novel in the sense that never before has a researcher thought that x was relevant topic.’
- They’re not paying you to write this review, and you enjoy anonymity, so let yourself go. Writing scientific article is such a dry, formulaic task – let your imagination run wild. It is a good way to unwind a little.
- It is always sensible to suggest that the authors cite some of your work, or more of it if they already have. The fact that the editor has invited you to review is basically only because they want your articles to liven up the drab submission under review. You may feel that suggesting reference to your articles means that you have to accept paper. Not so. As the authors pursue publication elsewhere, they will add your citations in any case, expecting you to review the paper again. If you are reviewing for a journal which is of a higher status than you yourself are used to publishing in, and if your work is cited in the article, you may wish to consider recommending acceptance somewhat automatically.
- It is standard practice to ask to reduce the length of the manuscript. NEVER say where the manuscript should be reduced, because for that level of contribution, you should expect to be an author on the paper. Instead, suggest some things to add in too.
- Always use dogmatic language. ‘The authors should add reference to the theory of adaptive loathing…’. This is more important for suggestions which are not justified or explained. It is more authoritative.
- Never use the first person or hedgers in your evaluations. The worst reviews start with, ‘In my opinion,’ ‘I found this article’, or ‘I think that….’ In general, the review should be written in the passive voice, to give it more authority and certainty, ‘This article is one of the …’. This authoritative tone is critical, because you will probably only want to review papers in areas for which you only have tangential expertise. You should refuse to review papers directly in your area on the grounds that your expertise may bias the publication process.
- On the other hand, you may use the first person as a useful dramatic device. ‘I was surprised not to find …’ is much more aggressive than ‘I think the authors might want to reference …’ Likewise, ‘I really struggle with the clarity of your results section’ is much more authoritative and emotive than, ‘the authors should read back the results and clarify x, y and z’.
- You are the reviewer and you have feelings. It is completely appropriate to use emotive and irascible language. It is only the authors who should not respond emotionally. Remember, there may be a graduate student author on the article, and they need to be taught a lesson in how the world works.
- The editor will not read the article, neither will they pay much attention to your comments, so make sure you are clear about your recommendations in the confidential comments to editor. The editor primarily uses these comments and the ratings to make their decision. Of course, your task is to make the editor’s decision for him or her. If your comment to the editor is at odds with the comments to the authors, its no big deal.
- Copy and paste from abstracts, and quote from your own papers. The authors appreciate that. If you have reviewed similar work before, you may use the same review again. You may even do this for a new version of the same manuscript. It just saves time, and as you know, timely publishing is much more important for research that scientific conduct.
- Never offer a full critique of problems with the article. You may get it back to review again, if you’ve not been sufficiently persuasive in your initial comments. In this case, you will want to have some new major problem up your sleeve to start afresh with your critique.
- Always be opaque. The editor is going to need to make a decision about the fate of the paper, based on the status, nationality, and track-record of the senior author, plus the published acceptance rate for the journal. Think of the poor editor who needs to write ‘unfortunately at our prestigious journal, we can only accept 10% of all papers which meet our exacting standards’. You need to make his/her task easy. If you offer a concrete set of problems which could actually be addressed, it will make this task very difficult for the editor. Better to make value judgements based on ‘feelings’.
- Finally, to fulfill your role to the best of your abilities, set yourself the task or remembering each and every occasion on which you have been rejected. Not just articles, but lovers, parents and classmates too. Think of your whole life and all the setbacks and unkindness that has ever befallen you. This will help you bring the right level of scientific diligence and clarity to the task. Remember, each paper that gets accepted for publication is one less paper on the Web of Knoweldge written by you .