44

Most of the time, when I am invited to review a paper, I am only provided with the title and abstract. The full manuscript becomes available only after I have agreed to review the paper. As per this question, this seems not to be common in all fields, but it does seem to be the norm in the journals I usually review for.

I assume this is a way to minimize the number of people that get access to the content of the paper before publication, and only provide that "privilege" to those who actually give something back in the form of a review.

However, it might hypothetically happen that I agree to review a paper based on the abstract, but then I see the full manuscript and find that I do not have the required level of expertise to review the methods used in the paper.

If I now decline the review, then I am going back on my word, and the community will see me as less reliable. Furthermore, I have unfairly gained access to unpublished content without giving anything back.

If I do the review, I will not be able to accurately evaluate the contributions of the paper, and might (probably) recommend acceptance of something that might not warrant it.

What is the best course of action in this case?

  • 15
    For the future, it's totally acceptable to ask the editor for the full manuscript before you agree to review it, so that you can make an informed decision. For the hypothetical scenario, I would say it depends: Did the authors misrepresent the used methods in the abstract? – lighthouse keeper Dec 12 '19 at 8:39
  • 7
    Is there a reason why you cannot briefly explain that from the title and abstract you thought the paper primarily involved "Method X applied to Topic Y" (or some such reasonably specific description), but instead you found that the paper involved [ ___ ] (fill in the blank with something reasonably specific)? In fact, if it were me, I would probably suggest a revision of the title and/or abstract that more clearly indicates the paper does NOT primarily involve "Method X applied to Topic Y". – Dave L Renfro Dec 12 '19 at 8:44
  • 2
    The canon in mathematics is to show the manuscript upfront. I rarely accept to review a paper without seeing it first, unless it is a request from someone I know very well. Since we get too many papers to review it is quite rare to engage in any type of discussion in order to see that particular paper upfront, it is a problem of the editor. – John B Dec 13 '19 at 1:33
  • @JohnB One reason this works is that it is very common that submitted math papers are already publicly available on arXiv, so restricting access to content isn't a consideration in the way that it may be in other fields. – Especially Lime Dec 13 '19 at 9:55
  • 2
    @TomášZato-ReinstateMonica No, you misunderstood. Reviewers (sometimes) agree to review seeing only the abstract but they review the paper reading the whole manuscript. – fa__ Dec 13 '19 at 13:37
101

The best course is to just inform the editor that you are unable to perform/complete the review as it is outside your area of expertise. The title led you to think otherwise. You are sorry, but others are better qualified to judge and give advice.

I don't see that as anything strange at all, and no one is likely to think less of you.

What would be wrong is to give a review of something you don't comprehend completely. No one benefits from that.

Just be honest with the editor. No one knows everything.

  • 54
    In fact, if the title/abstract is misleading, that is already valuable feedback. – ObscureOwl Dec 12 '19 at 11:47
  • 1
    In a couple of occasions I have added my comments anyway but being very clear of the situation and asking for a referee to be assigned. – Alchimista Dec 13 '19 at 7:58
29

Having very briefly -- and foolishly -- been an editor, I wish all reviewers shared your concern of "going back on my word". Far too many not only don't deliver promised reviews, but do so passively-aggressively by just not saying or doing anything, leaving editors to chase them down.

I also wish reviewers who discover they are not qualified, or for any other reason not able, to do a review would politely back out and give information as to their reason so that the editors can redirect the paper more effectively to someone who can do a good job. Crappy, unqualified or cluelessly superficial reviews are a real problem.

With all of this in mind, I would think more rather than less of any reviewer who backs out for reasons such as you describe, thoughtfully and in a timely fashion.

As to the "access" issue, indeed authors reasonably expect their unpublished work not to be blasted out to an unknown selection of colleagues/competitors in their field. All reviewers should not let unpublished work received in confidence influence their own work. Of course, there is an unavoidable "what has been seen cannot be unseen" element, but I'd expect this be less of an issue when you have decided the paper is outside your ability to engage with in detail, and when the expectation is you should stop reading it, I'd say delete it once your refusal is a done deal, and try to forget what you have seen.

17

I fully agree with @Buffy's answer, but would like to a add two "middle-ground" possibilities.

I as well as some colleagues have been in a similar situation in the sense that an important part of the paper turned out to be outside our respective expertise, while the reviewer was fine with reviewing most of the paper (as they concluded from title and abstract).

In such cases, the reviewer has been telling the editor that they are fine reviewing most of the paper, but it turned out that a substantial part is about topic outside expertise and they suggest/ask permission to consult named colleague who's an expert on that topic (example: expert on the application area consults expert on statistical data analysis).

My experience is that the editors were happy with that suggestion.
I suspect that it is sometimes very hard work for editors to find particular combinations of expertise in one person (consider particular application area + particular measurement technique + data analysis/statistics all of them at a sufficient level of expertise to judge whether the authors have been working correctly).

(I've been in the position of asking to consult a colleague as well as in the position of being the colleague who was asked to provide further expertise.)


In case my level of expertise is "almost fine" for the critical topic, i.e. I only need additional expertise confirming/refuting my conclusions but I'm confident that I can extract all relevant information from the manuscript, I've been abstracting (like a minimal working example) the situation and discussed my questions with colleagues without revealing the paper. As this is like reading up the topic in literature or a textbook, the permission of the editor is not needed.

6

What is the best course of action in this case?

Ask for the full manuscript before agreeing to review. I don't believe such access is unfair (you needn't look too deep when deciding whether to agree), I do believe you've given something back (your consideration).

5

Having already read the paper, you might as well write a quick review on those points you can judge, and send it in along with a note to the editor saying you strongly recommend appointing another reviewer.

As you are surely at least close to the subject at hand, you likely know a few names to propose to the editor as additional reviewers.

  • This is both sensible advice and a suggestion of alternatives would be helpful to the journal editor. – Norman Gray Dec 14 '19 at 15:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.