I am currently writing my bachelor's thesis in Computer Science. When I have a sentence

It was shown in [Kir10, HZK09] that ...

does it make a difference if I write

It was shown in [HZK09, Kir10] that ...


  • I've just took a look at 3 PhD thesis of my institute: All 3 of them are sorted by year, then by name. However, the bachelor's thesis I looked at had no obvious order. – Martin Thoma Oct 26 '14 at 7:32

Some styles require multiple citations to be in alphabetical or numerical order. For example, APA style requires that citations be in the same order as they appear in the references list (that would be alphabetical order.) Other styles (e.g. the Council of Science Editors) require that they be given in chronological order.

You need to check with the style guide for the style that you're following to see what is expected. If you're writing a journal or conference proceedings article, the instructions to authors should tell you what is required. If you're writing a thesis or dissertation, there are likely to be institutional policies that specify a particular style.

In LaTeX there is a package called "citesort" that will sort your citations. There is also a "sort" option for the natbib package that does this job.

| improve this answer | |
  • Conversely some styles require the reference list to be in the order of appearance in the paper (Eg IEEE) – Lyndon White Oct 26 '14 at 15:49
  • There are two different questions here: 1. What order do the references appear in the reference list (IEEE has them appear in the order in which they first appear in the paper) and 2. What order do citations within the paper take. e.g. (Jones 1979; Smith 1978) or (Smith 1978; Jones 1979.) The original poster was asking about the order in multiple citations rather than the ordering of the reference list itself. LaTeX and associated packages have the ability to deal with both issues. – Brian Borchers Oct 26 '14 at 17:12

First thing to do, is to check to see if there is a style guide for citations with your university and your supervisor.

If there isn't one, make sure when you choose an in-text citation style - keep consistent.

At my university, when I was a student (until last year), the rule was to use the first one in your example, the reason was to show a citation to the most recent works relevant to what you are citing.

| improve this answer | |
  • 1
    A little follow-up: Do you know if BibTeX can do this automatically? – Martin Thoma Oct 25 '14 at 21:47
  • @moose I have never used BibTex, but Using BibTeX: a short guide from the University of Toronto provides some information about BibTex citing. – user21984 Oct 25 '14 at 22:04
  • @moose that depends on the style you are using. Easiest way: compile latex/lyx file with [HZK09, Kir10] then with [Kir10, HZK09] and check. – seteropere Oct 25 '14 at 23:41
  • @moose That's under the control of the style, but in my experience most styles will render multiple citations in the order given. That seems logical - it's much easier for you to rearrange the order to be alphabetical if required than for you to somehow undo whatever decision the style makes. – sapi Oct 26 '14 at 2:20
  • 1
    natbib has an option to sort the citations- this can make your life a lot easier... – Brian Borchers Oct 26 '14 at 3:58

If the citations are sorted in the "expected order", nobody will notice.

If they are in the opposite order, people may assume that they were intentionally put in that order as an indication that the first one is more relevant than the later one. At least some people will, like me.

Look at the papers and books you read in your own field, and see what conventions they follow.

| improve this answer | |

As was correctly pointed out by several other answers, various styleguides require a specific ordering.

Brief rant: In my opinion, this is extremely unfortunate, and I cannot avoid the feeling that the only reason for such requirements is that editors feel that's an aspect where they can show with little effort (both alphabetical and chronological ordering can be checked automatically!) that they do provide and enforce clear styleguide rules.

What I would like to add to the other answers is the explicit statement that, if there is no fixed rule imposed by anyone, you can intentionally use the effect described in Oswald Veblen's answer. In that case, do order your citations in a way that makes most sense with respect to their contents, for instance, by listing the most literal example of a claim first.

After all, there is no reason to actually follow an ordering scheme by year, or by author name, unless you are required to do so. It conveys no useful information:

  • The earliest or the most recent publication is by no means automatically the most significant one.
  • Ordering by author name or paper title is even more arbitrary.
  • Ordering the same way as the bibliography has two aspects:
    • The bibliography might be sorted based on the order of first references to each entry. In that case, the ordering in a multi-citation is essentially irrelevant, as the bibliography will always adapt to that. This helps finding bibliography entries when reading a paper without the help of any technology that prevents manually searching for bibliography items, and also, it is only reliable for as long as each bibliography entry is referenced only once.
    • The bibliography might be sorted based on another criterion inherent to each bibliography item, such as year, title, or author name. In this case, references in a multi-citation are probably scattered throughout the bibliography, anyway, and using the same ordering for the references still conveys no useful information, as described above.
| improve this answer | |

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.