How to Referee a Paper

These instructions are for a first-time referee. Experienced referees may want to go to the checklist at the end.

The first, middle, and last thing to remember is your audience. That audience is usually only one person -- the editor of the journal who is trying to make a decision on whether to publish the article and is looking to your report for substantive advice.

Do not concentrate on a possible second reader -- the author (or authors) of the paper. If the paper is accepted, the author will usually not see your report. If the paper is not accepted, the editor may send your report to the author or only send edited portions of it

The editor wants an unemotional, clearly organized report that convincingly demonstrates that (1) you understand the paper, (2) you have a clear opinion about the paper, and (3) you have convincing arguments for your opinion and recommendations.

All of these points should be in a single document (kept to a single page if possible). There are two possible additional documents. A letter explaining a conflict of interest relevant to the editorial decision. Such conflicts of interest include (a) a paper on the same subject you have submitted or are about to submit; (b) any of the authors who are relatives, close personal friends, current collaborators, or recent students, post-docs or professors; (c) simple bias (pro or con).

List of typos. Authors always appreciate a list of typographical errors (``typos'') made in the manuscripts. Typos include misspellings, mistakes in sentences, equations, figure captions, etc. A separate list is preferable to embedding these in the primary opinion to the editor.

Checklist for refereeing

1. Summary of paper's reason for existence. he first paragraph should briefly but precisely summarize the main point(s) of the paper. Often your summary will differ from that of the authors. (The editor will detect this). If you cannot understand the paper that is reason for rejection.
2. Your recommendations. In that same paragraph (if it is short) or in the next (if it is not), precisely state your recommendation(s) with brief summaries of the reasons.
3. Reasons for opinion. In subsequent paragraph(s) explain each of your reasons with sufficient detail that the editor, who is not necessarily knowledgeable about the area but who is scientifically literate, can understand the report.
4. Quality of figures. Figures and tables are a very important part of the paper. Often the data is reported only there. Accordingly, their success in transmitting information is crucial to the success of the article. A rough rule is that a figure and its caption should be self-explanatory. The reader should not have to become a detective in order to deduce the meaning of a figure or table. The editor will -- and the author(s) should -- appreciate specific recommendations for improving figures and tables.
5. Two special reasons for rejection. Even if the science of the article is correct you should reject the article if (i) the subject is such that few readers would turn to this journal to find such an article and (ii) the article does not contain a ``least publishable unit'' -- that is, there is too little new material to justify writing an article about it.
6. Summary of your opinion. To aid the editor it is good to summarize your opinion and central arguments at the end of the report. \bye