1

Is there a recommendation or an acceptable number for computer science topic that is going to be published in journal? Is more than 300 references considered too much? What is the recommended number of references and is there a formula that I can use for calculating it?

Avi
  • 365
  • 4
  • 7
  • while I don't have an answer, it is certainly a relevant question as I know journal editors do pay attention to this. Helpful to look at your citations and see if there are old/foundational citations that are perhaps not directly relevant. Here is a good discussion on the topic: https://academia.stackexchange.com/q/13570/172246 – realkevlar Sep 20 '23 at 17:12
  • I once wrote an 80-page long survey paper (in math). It had 228 references, all essential for the survey. I would not consider 300 excessive for a survey paper if they are all essential. – Moishe Kohan Sep 21 '23 at 15:47

2 Answers2

4

No, there isn't any generally recommended number and no magic formula. I agree that >300 looks excessive, however this is not because there is a "right" number, but rather because a good survey paper should do more than just list references, so I wonder how long the whole thing is if you want to do all these references some justice. I mean it's clear that you won't discuss all of your references in detail; still you don't list references as a tick box exercise, rather you should list them because they have some relevance to the topic, and normally you should give your reader an idea what this relevance is because otherwise your survey would be pointless. It isn't enough (in the sense that it won't give the reader much of interest) to just note the existence of a paper.

The major rule here really is that you should always think about what's good for the reader, what will be interesting, what are references worthwhile to consider for the reader, and what will the reader learn about them from your paper. If there are really >300 papers worthwhile to mention (which is a possibility given your topic), chances are you chose your topic too wide. Also ask yourself, are all of these papers really worthwhile mentioning, in terms of impact and/or originality and/or quality? (If you don't have anything to say about them that may well not be the case.)

Ultimately you should cite all that's worthwhile citing, but your task is also to make clear to the reader (and yourself) why all these are worthwhile citing. It's not the number that is the issue; if your topic asks for that many papers, the problem may be the topic (and you should rather write a book).

That said, I'm not even ruling out that in certain circumstances it makes sense in that it may fulfill a certain purpose in a survey paper to just list a large number of papers without commenting more than a few words on each of them. If you have very good reasons you can probably convince editor and reviewers and it may go through. I'm sure I have seen survey papers well north of 150 references if probably not 200. As I said, the number isn't the issue, rather make sure you can convince yourself and your readers that all these deserve to be mentioned and it is good for the reader (who always has limited time) to do that.

Christian Hennig
  • 10,505
  • 22
  • 48
0

300 references is certainly excessive for anything other than a book. Other than that, it very much depends on what kind of journal and kind of article you are trying to publish.

A good approach to figuring out what is acceptable is to simply look through the articles you are citing. If all of them cite 10-25 other references, then that is what you should be shooting for as well -- so there's the "formula" you are looking for. What is "correct" is oftentimes simply determined by what is "common" in other examples of the thing you are trying to produce.

Wolfgang Bangerth
  • 94,697
  • 7
  • 201
  • 338
  • 4
    The title says it's a survey paper. A survey paper typically has quite a bit more references than the papers it cites. – Christian Hennig Sep 20 '23 at 20:01
  • @ChristianHennig True. But few will have 300. That will particularly be the case for survey papers written by people who are apparently inexperienced with these sorts of papers -- I suspect they will not be given 100 pages to survey a field. – Wolfgang Bangerth Sep 21 '23 at 03:14
  • I once had 228 references in an 80 page long survey. I would not regard 300 as excessive. – Moishe Kohan Sep 21 '23 at 16:19
  • Is this answer and the follow-up comment written specifically about CS? If they're intended as general statements, I think they don't account for the significant topic- and field-dependence in reference density (i.e. #references/#pages). – Anyon Sep 21 '23 at 18:29
  • 1
    @MoisheKohan an 80 page survey is well beyond what most people consider a "paper". Usually survey papers still fit within the usual page limits of papers, which vary somewhat among fields but are always well below 80. – Andrea Sep 21 '23 at 19:21
  • @Andrea: In math, this length is far from exceptional. For myself: In the last 6 years I had three papers which were over 90 pages long, published in a leading journal in my subfield. And these were not survey papers. – Moishe Kohan Sep 21 '23 at 19:33
  • @MoisheKohan I don't dispute that there are review articles that long and with that many references. But you probably didn't write yours at a time in your career when you were unsure about what number of references was appropriate -- people at that stage of their careers are not usually afforded 80 pages of space. – Wolfgang Bangerth Sep 21 '23 at 19:43