How To: Use Citations in a Paper

As you move forward in the class, we start emphasizing citing sources in more earnest. Most of us have had to do that at some point in our academic careers, but the reason we do this may not have been fully explained. Oftentimes it may have just been a requirement, like, “You must cite 5 sources”. If you were like I was, that meant I would write my paper, then go find five sources that were about similar topics and pretend like I had read them.

So, you may initially think this is just a collection of arbitrary requirements, so I want to take a moment to explain how you should use citations in a paper, which brings with it an explanation of why.

I generally identify five functional roles that citations play in a paper.

Role #1: Set Up the Fundamental Axioms of your Work

The first role is: to set up the fundamental axioms of your work. This means to set up the underlying assumptions and beliefs on which your work resides. That way, if someone disagrees with your work because of assumptions it makes rather than on what it does, it diverts that disagreement to the research that underlies your assumptions.

Imagine, for example, we want to do some research on homework usage in elementary school. So, we set up our argument like this: “Homework completion has no relationship with overall success at the elementary school level. Despite this, parents continue to judge the quality of education in part by the homework requirements placed on their children. In this research, we examine the effectiveness of a training curriculum to help educate parents on the lack of evidence for a causal relationship between homework performance and overall academic success at the elementary school level.”

If that’s how we position our research, then there are two bodies of people who are going to jump on us. Many people are going to say, “That’s wrong, homework is beneficial in elementary school!” Others are going to say, “That’s wrong, parents don’t think homework is valuable in the first place!” If you disagree with this work’s fundamental assumptions, you likely aren’t even going to bother reading the rest of the paper because it examines a phenomenon you don’t think exists.

Now, imagine if instead, we wrote this: “Two comprehensive meta-analyses (Cooper, Robinson & Patall 2006; Fan et al. 2017) revealed little research to support a link between homework completion and overall success in elementary school. Others have suggested that the value of homework may be as an anchor of parental involvement in education (Domina 2005), but there is little evidence that parental involvement is diminished if homework is removed. Despite this, parents believe that homework is valuable (Cunha et al. 2015; Valle et al. 2015). In this research, we examine the effectiveness of a training curriculum to help educate parents on the lack of evidence for a causal relationship between homework performance and overall academic success at the elementary school level.”

With this presentation, if you disagree with the axioms of our research, then it’s clear your disagreement is with Cooper, Robinson & Patall, Fen et al., Domina, Cunha et al., and Valle et al.. This work, however, would clearly have a sufficient theoretical foundation on which to build.

That’s not to say that having something to cite always gives enough foundation: there are times when later research refutes earlier research, in which case the criticism is, “The axioms on which you are building have been refuted by other research, and so the responsibility is back on you to defend the ones you chose.” But generally, citing research shows that you are building on solid theoretical ground.

Role #2: Demonstrate Your Awareness of the Big Picture

Second, a major goal is to contextualize your work within a broader picture. To explain why that’s valuable, here’s an image:

In science, we’re generally all looking at some underlying phenomenon from different angles, and it is the combination of all those angles that help us uncover the real truth. Our research is the equivalent of the people in the cartoon above getting together to discuss what they each found, and coming to the conclusion of, “Oh, it must be an elephant, that explains how it could simultaneously be a tree trunk, a rope, and a spear”.

Citing research demonstrates that you’re aware of what others have found about that same underlying truth. A major weakness in many people that come into research is a belief that they know better than everything that has come before, and therefore they need not bother with learning what others have done. That ends up being a surefire way to make sure no one listens to your work or findings: it shows that you don’t know enough about the big picture to make the claims you’re making. It’s the equivalent of someone new walking up, feeling the elephant’s toes, and saying, “Hey, it’s a rock!” and insisting without listening to what the others have found. By citing the research, you’re saying, “I’m aware of the prior research that suggests this is an elephant. However, all prior research is also compatible with this being a rhinoceros, and I have new evidence that is only compatible with the Rhinoceros Hypothesis.”

Role #3: Contextualizes Your Work in the Big Picture

The flip side of that is that it lets you also direct attention to where your contribution is. Imagine in the above example, you’ve come along and observed that the pointy widens to a wider base, and the snake and rope are actually the same thing — both ideas that would support that the elephant is actually a rhinoceros. Citing research lets you clearly state, “In the big picture of this research, we have uncovered evidence that specifically casts doubt on these two portions of what we believed in the past.” It clarifies where your contribution lies to make it easier to see the big picture. Without that, it might come across more as, “There’s also a horn!”, leading to a revised image of an animal that has both a trunk, two tusks, and a horn.

To bring that around to our general domain, consider a paper I have at Learning @ Scale this year. The paper shows that in our online CS1 class, students learn as much as they do in a traditional class. However, there is significant research that shows that students in online classes do not learn as much as students in traditional classes. Without describing the big picture, my paper would merely be shouting “Yes they do!” back at a crowd shouting “No they don’t!” But by citing the research, we are able to help build a bigger picture. We basically say, “This research that shows students learn less in online classes was conducted in small community colleges with few resources to develop strong online offerings. Our research was conducted in a research institution that put considerable resources behind this development. Therefore, it may be that online is not inherently better or worse, but that there are execution or resource details that determine its success.”

Role #4: Sets Up the Breadcrumbs to Past and Future Research

Fourth, citations help build a network that people use to navigate a literature history. Citing your research provides the intellectual genealogy of your work, allowing those that follow the field in general to find it, and to understand its relationship with other ongoing efforts.

This connects back to our previous topic on how to find papers to read: the citations help someone navigating your field peruse the work that led to it, leading to a greater understanding. Modern tools also help make citations two-way: you can find pages like this that list the papers that went on to cite the current paper, navigating both forward and backward.

Role #5: Diminish Suspicions of Plagiarism

Fifth, plagiarism is not only an issue in education: it is also an issue in research. It is sadly not uncommon for some researchers to blatantly plagiarize. More commonly, though, plagiarism exists in the form of overstating one’s own claims by tacitly incorporating the work of others into your own. For example, in the paper I referenced above that will be at Learning @ Scale this year, I posited the idea that the effectiveness of an online class is in part on the resources (human, financial, and technological) used to create and support the individual online class. Someone else working in the same area could come along and pose the same model as if it is original, knowing that there is increasing evidence that that dynamic exists. Such disagreements are the foundation of famous arguments like the Leibniz-Newton controversy, where scientists argue over whether two scientists could truly come to the same model independently, or if one had to have been aware of the other’s work.

This has both gotten more and less difficult in recent years. Technological advancements mean that it is possible to communicate much faster, but that means that the quantity of communication has increased. It is common for scientists to independently propose similar ideas because they are not aware of one another’s work. The question then becomes: did you do your due diligence in looking at existing research to determine if this was an existing idea?

Here’s another example of this from my own work. As part of my dissertation work, I developed a little typology of the types of evidence people use to support their hypotheses: direct observations, analogies to other systems, testimony from experts, etc. Nothing like it existed in cognitive science, artificial intelligence, or learning sciences literature that I knew of. I later learned that there did exist a similar idea in epistemic cognition from only a couple years earlier. Here, it was reasonable to believe that I was truly not aware of that other theory because it was not within my academic community; and, once aware of it, I specifically referenced it whenever bringing up my typology again.

Citations show that you did your due diligence in ensuring your research is original.

Other Roles

These are, in my view, the intellectual functions of citations. That said, there are other roles they play as well. One of the significant roles I believe they play is in more generally demonstrating due diligence. To draw a crude analogy, there is an old adage that if you want to know how clean a restaurant keeps its kitchen, check its bathroom. A restaurant that does not clean its bathroom very well likely doesn’t clean its kitchen well, either. For a paper, if you want to see the amount of thought that went into the work, check how much went into the literature review. If the literature review is phoned in (as many are), chances are other components of the research are as well. I believe having a solid literature review institutes a sort of Halo effect: the reader, impressed by the depth of the related work section, generally gives the paper the benefit of the doubt in other ways. “After all, if they put this much work into reviewing the literature, they likely put enough work in for there to be answers to my other concerns, too.”

And lastly, I’d be remiss if I didn’t acknowledge that part of the role of citations is to adequately “bow” to the established figures of the field. One goal of a comprehensive literature review is to raise the likelihood that you cite the people who end up reviewing your own paper and deciding whether it’s accepted. I have gotten several reviews where it seemed clear (although it isn’t provable) that the reviewer’s major complaint was that we didn’t cite them, and ended up nitpicking in other places as a result (sort of a reverse-Halo effect).

TL;DR:

So, why do we cite related research? It’s not just to check off a requirement on a rubric. It’s to put our work into a greater context, prove we have the foundation to be able to contribute to that growing area, and show that our work is original.