Nowadays data set providers typically ask data set users to cite their paper(s) introducing the data set. When did this practice become commonplace?

(Given what I read on this page, I feel the need to emphasize that I do not intend to open a debate around the usefulness of such citations.)

  • Since when has it become common to write papers based on someone else's data set?
    – keshlam
    Jul 12, 2015 at 3:16
  • 10
    @keshlam Since people share data sets? Jul 12, 2015 at 4:45
  • 4
    In that case, my answer would be the same. If someone has made a major contribution to your paper, you owe them recognition... and you owe your readers the ability to trace the data back, in case there was something unusual about it. Credit or citation would seem the obvious ways to do that. I'm not sure why you believe this wouldn't have been done in the past; i'd think it was simply "best practice".
    – keshlam
    Jul 12, 2015 at 5:44
  • 1
    @O.R.Mapper I am asking since when data set providers ask users to cite their paper(s) introducing the data set, i.e. one or several particular papers that, in the data set provider's opinion, is most preferrable to represent the dataset. Jul 12, 2015 at 16:38
  • 1
    Question to the OP: Is the emphasis in your question on the "issue" that the authors of a data set ask a certain paper to be cited instead of the dataset itself? Or is rather that you are interested in from when onwards they ask for to get any type of citation at all?
    – DCTLib
    Jul 13, 2015 at 8:31

1 Answer 1


I don't see a logical distinction in citing used algorithms vis-a-vis data sets. Both seem perfectly natural to get cited. Both are an integral part in the results you are reporting in a paper and hence should be attributed.

You're probably thinking in terms of machine learning papers, where typically a new method has to be benchmarked on a large variety of data sets. While it may seem like overkill to cite each and every data set that was used, lets not forget that whoever provided that data also put in effort and deserves credit for it. Again, this isn't that much different from citing the competing algorithms you are comparing against.

(Academic) software packages should also get cited, or at least mentioned, when they are used. This isn't really a new thing, for example the old 4-clause BSD license essentially demanded the same thing.

  • 2
    @O.R.Mapper I know that happens often, but both are bad practice in my opinion. Jul 12, 2015 at 10:24
  • 1
    I fully agree. It is just that this answer (and also some of the comments on the question) make it sound like we had to explain to the OP why datasets (and software, and similar) should be cited (which may or may not be necessary to explain), while the explicitly asked question, when dataset providers started to exlicitly suggest particular papers on the dataset for citation, is not really answered. Jul 12, 2015 at 10:41
  • 1
    @O.R.Mapper as far as I know there is no when, it was always the right thing to do. I think that's why the comments and my answer are what they are. Jul 12, 2015 at 10:59
  • 1
    Again, the question as it is written is not when it became the right thing to do, but when it became common practice to explicitly write something like "If you use this dataset, please cite paper XYZ." on the website from where the dataset can be obtained, rather than relying on the authors to do the right thing. Did the very first authors publishing datasets already do so? Or were they rather completely happy with providing this novel service to the community by providing a dataset, whereas the idea that you could actively ask for a citation thereof evolved only at a later time? Jul 12, 2015 at 11:11
  • 1
    Since Fisher's Iris data set had an accompanying paper I guess it has always been quite common to get cited for data. I reckon Iris is one of the first heavily reused data sets. Even if it wasn't explicitly asked in the past, source attribution has always been important in academic writing. Jul 12, 2015 at 11:18

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .