The Art of Clean References

My guidelines on how to maintain clean references in academic writing.

Note: This is a new import of an old blog post from 2018.

There are so many factors that are out of your control when you submit a paper for peer review, so my philosophy is to own the factors that you do control. One of them is your list of references. Having “clean” references — references that are well-formatted, complete, and correct — is a hallmark of quality for a paper in my eyes, yet I see so many submissions (even from well-known and experienced authors) where the references are a shambles.

In this document, I will talk about what I consider to be “clean” references and how to check your submissions to conform to this standard. As is often the case for my guides, this one was primarily written for my own students to refer to when working on a paper with me. Whenever I found myself reiterating the same thing to my students more than a few times, I would write one of these documents and put it on my research group’s internal Wiki. I wrote the first version of this particular document on references in September 2009.

With this background, let me also add the caveat that I do not claim that any of these “rules” are universal or even best practices; they merely represent the standard I use myself for academic writing. I only ask the students who co-author work with me to follow this standard. In other words, unless you are writing a paper with me, you can safely ignore these rules. Nevertheless, I believe there is useful information here, and I urge everyone to spend some effort to keep their references clean. Why? Keep reading.

Motivation

Why is clean references such a big deal? Well, as I stated above, the first reason is that I, personally, see clean references as a sign of quality in a paper. Messy references with poor formatting and incomplete information, such as missing authors, page numbers, and years, indicate that the authors were in a hurry to complete their work. This detracts from my perception of the submission, and I surely can’t be the only reviewer to feel this way.

I’m also pedantic to a fault, which I think is a good trait for an academic. I can’t stand messy references in my own submissions, and when I see it in a paper I am reviewing, it insults my eyes and my sense of aesthetics.

Finally, you never know who your reviewers will be when you submit your paper. While not citing a relevant paper the reviewer wrote may be a mortal sin, citing that paper incorrectly is just one step removed. For example, people tend to be sensitive about their names, so misspelling a name in your reference list can have a negative impact if that person happens to be reviewing your paper. As a case in point, my own first name is actually “Niklas” and not “Nicholas” or even “Nikolas”, and my last name has a “q” followed by a “v”, not “Elmquist” as I so often see it misspelled (all of these are examples of Swedish spelling, which I am proud of). This is not to say that I would ever reject a paper for the “impertinence” of misspelling my name, but it is a small annoyance that can add up to an overall negative perception of your paper’s quality.

Even for students who are working with me, the penalty for not conforming to this standard is pretty mild. All it means is that I am going to be spending all of my time fixing their messy references until they are clean rather than working on the body of the paper. Most students I know would rather prefer that their professor spends his or her valuable time on the actual writing rather on the mundane task of keeping the references accurate. This is precisely the message I want to send: references may be mundane, and anyone with an internet connection can update them, but that doesn’t mean they are not important. They are.

The Rules

First of all, use a reference management system, even if it is only BibTeX (which is actually what I use; I never got in the habit of doing anything more fancy). Managing your references manually, especially when writing in something like Microsoft Word, can be very painful. Pick a reference management system and learn how to use it competently.

Second, you do not always have to write your references yourself. There are some good databases on the internet (for example, for BibTeX there is the The Collection of Computer Science Bibliographies ). Always look in a database first to see if your paper has already been included in the database, in which case you can just cut and paste (of course, these are often dirty, so you may still need to clean them).

Here follows my informal rules for clean references:

• Consistency: Above all, while you are free to ignore the guidelines below, at least be consistent. You may feel it your personal prerogative to abbreviate “Proceedings of the ACM Conference on Human Factors in Computing Systems” to “Proceedings of ACM CHI” or even “Proc. CHI”, but at least use the same treatment for all of your references. Otherwise, it looks sloppy and careless.
• Accuracy: Always endeavor to include as much information as possible in the references even if the final output format does not require it. For example, some reference formats will abbreviate all first names to their first character, but this is no reason to avoid adding those full first names to your reference database. The next paper you write may require full first names.
• Completeness: Make sure that you maintain the minimum of information necessary for a citation. For a conference paper, you will need to make sure that you have, beyond author names and paper title, also the name of the proceedings, the year, and the page numbers. For a journal paper, you will need the journal name, the volume, the number, and the page numbers. Publisher information and address can also be informative. If you do not have some of this information, you better start looking it up.
• Conference names: A common mistake that I see is that people are sloppy about their conference names. Remember that we are naming the proceedings, not the conference itself. In other words, you should not use “IEEE VAST’09” as the proceedings, because that it is incomplete and incorrect. The full name of the proceedings for a paper in VAST 2009 would be “Proceedings of the IEEE Symposium on Visual Analytics Science and Technology”. Remember what the conferences actually are (symposium, workshop, or conference).
• Avoid abbreviations: I generally avoid including the abbreviation of the conference or journal in the reference because I find it to be redundant. In other words, I don’t write “Proceedings of the ACM CHI Conference on Human Factors in Computing Systems”, but instead just “Proceedings of the ACM Conference on Human Factors in Computing Systems”.
• Avoid years in the publication name: Do not include year numbers in the name of the conference proceedings. There is typically a specific field for year in a reference entry, use that.
• Journal versus conference papers: Please remember that some conference papers are published in special issues of journal. To cite these, cite them as journal papers (including volume and number), not as conference papers. Just eliminate the name of the conference from the reference entry.
• Digital Object Identifier (DOI): The DOI is a way to uniquely identify digital objects, and most scientific papers have unique DOIs. Contemporary conference and journal paper formats now tend to include DOIs, so I try to include them in my reference entries as well. Remember to be consistent here: if you include it for one reference, be sure to include it for all your entries (except for things like workshop or poster papers, which may not have a DOI).

Below are some specific rules for working with BibTeX, which I include here because it is what I tend to use:

• Publication types: All BibTeX entries have a type — -this is the type of publication you are citing. For conference papers, it will be @InProceedings, and for journal papers, it will be @Article. Keep these straight, and also remember that there are other types if you want to cite a book, a chapter in a book, a tech report, a Ph.D. dissertation, etc.
• Author names: Remember that author names in .bib files are NOT delimited by commas, but by “and”:s. In other words, if you have three authors, you would write them as “Andrew Almond and Buster Banana and Chester Cashew”.
• BibTeX output: Read the output from ‘bibtex’ when you run it on your .bib file. Many times it will tell you if you have multiply-defined labels, or if there is something wrong with a particular BibTeX entry. Pay attention and go back to fix those mistakes.

Conclusion

Reference cleanliness is next to godliness. Spending the time to put your references in order is one small thing you can control yourself, and can communicate much about the quality and maturity of your submission (at least if I am your reviewer or advisor).