In a publish-or-perish system of scientific communication, large numbers of articles are written and submitted to journals. Editors face the challenge of finding peer reviewers for all these articles. Scientists who accept invitations to review, usually for free, play a vital role in the system. This gives them a lot of freedom to choose which invitations to accept, and to perform peer review as they wish.
This blog post states my own current “code of conduct” as a peer reviewer: my answers to questions that arise whenever a researcher receives an invitation to review, or has an article to referee.
Should I review this article?
1. Only review for journals I would publish in. This means the publisher should not be Elsevier, the journal should be open access, should not have obnoxious formatting constraints, should not cost too much to the community, and should publish reviewer reports.
2. See the article’s full text before deciding. And the article should already be available on arXiv as a preprint.
3. The article should be clear enough that I can concentrate on the contents, and not waste too much time on issues of style and presentation. Unclear articles are sometimes sent to peer reviewers, while they should in principle be desk-rejected by editors.
4. The article should be really interesting. There is no point in working on an article if I will be the only person who ever reads it! Some articles are written for inflating the authors’ bibliometric statistics, others do generalizations for generalization’s sake: no one should waste time on such articles. Exceptions can be made for articles by PhD students, if the reviewer’s work can help them learn the trade.
Working on the article
5. Directly contact the authors whenever needed. When reading an article, it can be helpful to get clarifications from the authors by email or other channels. There is no good reason to renounce this when reviewing for a journal.
6. Do it early, do it quickly, in one go. I usually devote of the order of 1-2 working days to reviewing an article. Interactions with authors are more productive if the article is still fresh in their minds.
7. Read the article in full, while paying more attention to the important parts. In decreasing order of importance: title, abstract, introduction, main text, appendices.
Writing the report
8. Write for the public, then for the authors. The report should help other scientists decide whether and why the article is worth reading. It should also help the authors improve their text, and possibly their research.
9. Write an introduction to the work, then a numbered list of questions and suggestions. The introduction states why the article is interesting, and to what extent the results can be trusted.
10. No self-censorship of suggestions. Suggest improvements of the form or contents of the article, without regard for the needed amount of work. The authors will decide which suggestions they want to implement.
Sending the report
11. No explicit recommendation to publish or not. This also means I happily ignore the journal’s criterions and demands to reviewers. If forced to make a recommendation, I may add a disclaimer of the type: “As a matter of principle I do not make recommendations on whether to publish articles or not. Editorial decisions belong to editors. Since the computer system makes it necessary to have a recommendation, I picked the first one in the list.”
12. Sign the report. Reviewer anonymity can lead to reports that are weak or biased.
13. Immediately send the report to the authors. Better discuss with the authors right after studying their work, than weeks later through the journal’s system.
14. Follow-up reports are generally cursory. Journals sometimes ask follow-up reports on amended versions of the article. A typical follow-up report says that the article has been improved, and that I have no further suggestions.
Examples of past reports
On SciPost: 2025, 2024, 2024, 2023, 2022, 2022, 2018.
On Wikiversity: 4 reports are accessible from this page.
No comments:
Post a Comment