How to Choose Citations

Choosing papers to cite is a key step in preparing a manuscript, but you shouldn’t take it lightly. Consider that a highly cited paper is often assumed to be a great paper. I don’t believe that this should be automatic, but I recognize the value of this (relatively) objective metric for judging someone’s work. With this in mind, this post offers some general tips on how to choose citations.

I hope that this post isn’t groundbreaking, but I wanted to establish a framework that I will use in future posts on citations. This subject is one that I have reflected on considerably, both as an author and as a reviewer, and was one of my primary motivators to begin Rambling Academic. I’ve had the privilege of reviewing 41 unique manuscripts (mostly journal articles) as of the end of 2015, and problems with references has been a common theme.

I hope that this post will help guide researchers who are relatively new to writing or reviewing manuscripts. As an Engineer-in-Training, I’ve recently come to appreciate how professionalism and ethics in engineering can be applied to academic research. This has helped me to articulate my philosophy on choosing citations.

Types of Citations

I place references into 1 of 3 broad categories:

1. Specific References. This is the “easy” category. If a paper uses an equation, definition, claim, quote, etc., from another paper, then you should cite it. There’s a limited amount of flexibility here, but still opportunities to falter.
2. Foundation References. This is another somewhat “easy” category. If your topic of research can be traced back to a particular seminal work, then it is often expected to cite that work. Common examples could include a survey paper, a textbook, or an influential thesis. I don’t believe that you have to cite a work because everyone else is doing so, but at least one or two foundation references should be included from somewhere.
3. General References. This category is much more open and therefore more challenging. Papers in this category would be classified as “related work”, and don’t necessarily have a direct connection to the current manuscript. These papers are often mentioned once and then never referred to again. There is an incredible degree of flexibility here. I believe that this discretion also demands responsibility for you to make your selections carefully.

Guidelines on How to Choose Citations

I’ve tried to base my citation philosophy on accuracy, fairness, professionalism, and a positive attitude. With those in mind, I offer the following guidelines on how to choose citations:

1. Avoid Conflict of Interest (Actual or Perceived). Conflict of interest in this context is best shown by self-citation. I think that your default approach should be to not cite you own work or the work of your close collaborators (whether or not they are a co-author on your current manuscript). Any time you do so, you create a conflict of interest. Even if the inclusion of a self-citation is justified, there is at least a perceived conflict of interest. The threshold for including a self-citation should be higher than any other citation, especially for general references. I list this guideline first because I often see self-citation as part of the violation of any other guideline. I discuss justified self-citations in another post.
2. Find the Original Source. Let’s say you cite paper A for a particular idea. If paper A only mentions that idea when citing paper B, then you should just cite paper B. If paper A has an extended discussion of the idea, and meaningfully expands upon the relevant discussion in paper B (perhaps by comparing with papers C and D), then citing just paper A is justified. Otherwise, you’re forcing the reader to look at paper A just to get to paper B. If paper A is a self-citation, then you really need to act carefully to justify its citation over paper B.
3. Be as Specific as Possible. This guideline is straightforward. If you’re citing specific content, then you should refer to that content directly. For example, provide an equation number if you’re citing an equation, or a chapter number if you’re citing a book.
4. Justify Every Citation as a Whole. Consider the relevance of a general citation as a whole. If there are only one or two relevant sentences, then you may want to find something more relevant. You should create a list of references that are as helpful as possible for the reader.
5. Be Positive. This is the last guideline but perhaps the most important. It’s tempting to take general references as an opportunity to criticize other people’s work. You should give other authors the benefit of the doubt and not try to promote yourself at their expense. I don’t say this just because one of those authors might be a reviewer. I say this because your list of general references should be a list of papers that you actually want the reader to read. Why cite papers that you wouldn’t recommend? Of course, you need to compare and contrast your contributions with other related work, but this should be done respectfully.

Is this a helpful list? Did I leave out an important guideline? Do I appear too strongly against self-citations? I have more on self-citations here.