Suppose I'm writing a paper and at some point in the paper (e.g. the background section), I write that "Vanilla ice cream is healthier than chocolate ice cream." I search EBSCO and find check all research studies that tested this and find and skim through 15 papers all supporting this claim, including the earliest pioneer research in the area up to the most recent. Is it then appropriate to affix in-text citations for all 15 of those studies to that sentence? What about if I find 50 studies?

  • 1
    Would you be writing "Vanilla ice cream is healthier than chocolate ice cream" before you had some reason to make that claim? Commented Nov 20, 2014 at 4:19

6 Answers 6


It depends on the purpose of the paper.

If it is a review article on the effects of different flavor of ice cream, it may well be appropriate to cite far more papers than it would be if you are merely establishing the point in a paper containing your original research on the subject. In the latter case you may even be able to cite someone else's review.

Generally speaking, new papers in a field only get published if there is an element of novelty to them. Thus it is unlikely to find 50 different papers saying exactly the same thing. If you need the citation to support your own project, you should be able to find a small subset of those that are most relevant to the question at hand (e.g. if I am using mouse models, I should probably cite prior work in mice over prior work in rabbit models). If you are reviewing a topic, you should be able to find the most important papers out of the group (e.g. the ones that were transformative to our understanding of ice cream health effects).

Additionally some papers will vary in quality or scope. Citing an initial, limited scope paper might be appropriate if you are discussing the history of a topic, but if you are just trying to establish a basis for a claim you may want to go with a later, more definitive study.

Many publications will give you a citation limit when writing for them. So you often won't have a chance to cite an entire library of previous work. You need to be able to pin down the papers that are best suited for your specific need.


The purpose of a citation is twofold:

  1. It prevents you from taking credit for other person's work.
  2. It can serve as an "external appendix" (the interested reader can read more in ...)

Neither requires that you cite everybody. If there is a large body of work done already, then typically one or more overview articles have already appeared. You can cite those, and combine both purposes: you indicate that it is not your finding and the interested reader can find a complete list of references in the overview article.

  • In case there aren't overview articles, you can always cite a few you like, and say something like "see these and the references therein."
    – Kimball
    Commented Nov 20, 2014 at 8:31

In my field (particle physics), at least, when you give a citation for a result, it's traditional to cite the original result. Many papers that come later might also give the same result, some with the original derivation, some with updated derivations, some with experimental evidence (though that is really considered a separate result from the theoretical discovery in its own right), but they don't all get cited for that one result.

So it probably depends somewhat on the conventions of your field.

  • Of course, this practice also has its problems, especially when the original paper is very old, in an obscure journal and/or written in a language you don't understand. After all, you're also supposed to have read and (more or less) understood any articles you cite. Commented Nov 19, 2014 at 17:58
  • 2
    @IlmariKaronen I haven't read most of the papers where I am listed as a coauthor, let alone all of the papers that I put in the references section of my master's thesis. (Particle physics papers tend to list as coauthors everyone even tangentially related to the paper, in my very limited experience.) And I am as close to a nobody as you can be in the world of coauthors of particle physics papers. Commented Nov 19, 2014 at 20:59
  • It's mostly the large experimental collaborations which have long lists of authors that appear on every paper. In any case, I have gotten the sense people in particle physics don't always hold citations to that standard. For example if you cite an experimental result in a theory paper, the authors (who are theorists) would not be expected to understand the details of how the experimental apparatus works.
    – David Z
    Commented Nov 19, 2014 at 21:01
  • 1
    @yellowantphil: It may happen all too often, but most publication style codes do consider it improper to cite a paper none of the authors have ever read, at least unless the citation is clearly marked as indirect (e.g. "Lomonosov 1759, qtd. in Shiltsev 2014"). Commented Nov 19, 2014 at 21:26

It depends on how important is that point to your paper. If you are studying the consistency of ice cream, and you find that vanilla has a better consistency, and furthermore, is healthier, you just need one good reference.

On the other hand, if that is a central point of your article (you are researching how to condition children into choosing vanilla instead of chocolate), you should make a strong point. In this case, is beneficial to put things in perspective:

The first work by Jones in 1976 suggested that Chocolate was not so healthy...

If there is a review, it can mostly replace all the older papers:

The fact was well established since the work by Marks, "Ice cream flavours, health, and little humans", 1998.

and possibly special mention mention those papers that are a substantial breakthrough or strength:

Smith 2003 followed the ice cream preferences of 10 000 white children...


+1 to Maarten's answer.

In addition: likely enough, the more recent of the 50 studies you mention already cite most of the earlier ones, so the reader can just follow the breadcrumbs back - no need for you to essentially repeat the literature survey that was already done in the papers you cite. Stick (mostly) with the recent literature.

Of course, you should cite papers that are most relevant to your study (e.g., similar protocols, similar questions etc.), and if you manage to find a relevant study that everyone else has overlooked, by all means include it.

  • This makes sense, unless there's a seminal paper. A think a huge problem is that not enough people actually follow the breadcrumbs back before writing. Commented Nov 19, 2014 at 21:18

Usually one reference is best; sometimes you want more.

In general a single, well-chosen, citation is all that is needed. Select a paper that (a) you've read (b) clearly supports your assertion and (c) is well written and the job is done.

However, sometimes you may wish to include more than one references. The appropriate time to do this is if there is historical reason to do so (e.g. you may wish to cite the original paper and a more recent update), if you wish to emphasize the breadth of support for the point, if the point is - in your view - controversial or if you're giving a historical view and want to cite multiple papers to give the historical perspective on the point.

You must log in to answer this question.

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