28

At the end of writing up our manuscript, we discover that there has been another earlier study in a very different context (but using a physical model very similar to ours) that claims a similar result.

We were not inspired by this earlier work and we could not have found it until we had our results; so this is not a case of not having done proper literature survey a priori.

So, do we cite them within the main text's introduction and results despite not having been inspired by it at all? Or do we cite them during the concluding remarks highlighting the similarities?

In either case, our results certainly complement theirs.

36

In mathematics (maybe in other fields too), one would put at the end of the introduction a statement like "After obtaining the results in this paper, we learned of related work by X. In particular, X obtained ...." Here the "..." would be a description like "a stronger form of our Theorem 7" or "a weaker form of our Theorem 7" or "a result related to our Theorem 7" or whatever it was that X actually got.

In computer science (maybe in other fields too), papers often have a separate section called "Related Work", and information of this sort would naturally go into that section.

  • 1
    In other fields, this approach would be a kind of risky - reviewers might complain, unless you're writing for a conference and running short of time for the deadline. If you're writing for a journal, you have all the time to modify the paper, properly putting in perspective a work which is so closely related to yours. See also peer revu's answer. – Massimo Ortolano Mar 13 '15 at 19:20
  • This is also risky in math and CS. On the other hand, the alternative is to be dishonest. Risky or not, if you know of earlier work that claims the same or similar results, you have to cite it. – JeffE Mar 15 '15 at 12:17
  • 1
    @JeffE: What I considered risky is the idea of writing just an addedum at the end of the introduction: one should take the time to completely rewrite the introduction, if the article to be cited is so closely related to their own. – Massimo Ortolano Mar 15 '15 at 14:53
  • 1
    And, indeed, if necessary, also part of the other sections... – Massimo Ortolano Mar 15 '15 at 15:00
19

It is useful for readers if you distinguish your work from previous work early on in a paper, typically in a "related work" section. If your paper is being peer reviewed, distinguishing your work from previous work is an important aspect of demonstrating novelty.

It does not matter that you didn't find out about the previous work until after you had results. The manuscript is not a chronological record of your thought process.

  • 3
    Completely agree. The manuscript should be rewritten before submission to include the fruits of recent research. There is zero reason not to do so. – Scott Seidman Mar 13 '15 at 19:39
12

So, do we cite them within the main text despite not having been inspired by it at all?

Yes, you should surely cite it and you can do it in the introduction, where you will outline the differences between the two works and specify in which way yours complements theirs.

  • 2
    Could you explain why it is curious? The systems considered are really in very different contexts, but uses similar models and therefore gets similar results (the new result being a particular phenomenon); the earlier study highlights the new phenomenon in their title and abstract in the context of their totally different system. So only once we had the result could we have stumbled onto their paper. How common is this situation? – MaviPranav Mar 13 '15 at 13:06
  • 3
    We were writing the manuscript and simultaneously searching for similar previously published reports on such a phenomenon, and found this earlier work only at the end of writing. – MaviPranav Mar 13 '15 at 13:31
  • 5
    @MaviPranav It is not curious that you found the work very late the in the process (this happens all the time, for better or for worse). What is weird is that you say that you "couldn't" have found it earlier. Why not? Abstractly, I could imagine that you only knew after getting all your results that you knew what related work you need to look for, but I cannot really imagine a practical situation where this would be the case. – xLeitix Mar 13 '15 at 19:02
  • 5
    @xLeitix: I suspect this is just a language issue at work here. I have seen non-native English speakers from various cultures use constructions such as "could not have found" when they mean "did not find". The "could not" is not meant to express general impossibility, it expresses failure to achieve the described result. – O. R. Mapper Mar 13 '15 at 20:15
  • 4
    @xLeitix: No, I indeed meant "could not have found" but for practical purposes as described by anonymous_reviewer below. I was going to cite the work anyway as I had suggested in the post, but I was unsure whether their work and results should be alluded to right in the introduction --- which might give the disingenuous impression that our study was inspired by their work done in a very different setting ---, or a discussion of their results should be relegated to the end --- which would seem like a disingenuous downplay of the importance of their work ---, both of which I wanted to avoid. – MaviPranav Mar 14 '15 at 17:00
4

If indeed the other study deals with a completely different setting, and is only related to the present one by an analogous finding as the main result, and the only way to come across this other study is by using a description of that result as search terms, then I do not see why it should be cited in the introduction. Unless, of course, it is something to potentially attract further interest towards reading the present paper in full (the introduction is a good place to try to attract a potentially interested reader to read the paper completely).

Otherwise, something like this can be mentioned in the "discussion" part of the paper, whether it is a separate section or a part of the conclusions.

It is not like before beginning any study, we are first going to carefully read through all the papers in the world in order to find out if similar models have possibly been used somewhere in a completely unrelated setting. If the only reasonable way to come across the other study is indeed the result itself, it is fine to leave this citation to the end of the paper.

0

This also very much depends on the journal you submit to. There are a growing number of journals that aren't so concerned with the novelty of the result as compared to whether the science was technically correct. For example, PLoS One would likely accept the paper even if you note the other similar result. Some rarer journals may even value that you've independently found similar results as another.

protected by Community Mar 18 '15 at 3:48

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

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