This post is meant to serve as an introduction to the art of reading academic research articles for casual, non-expert readers such as undergraduate students. As such, it does not presuppose any knowledge of statistics or even the scientific method. The goal is to give the neophyte reader a set of rudimentary tools that can be implemented regardless of what research question a paper tackles. The hope is that readers will be able to extract from an article the kind of information they are looking for in an efficient and systematic manner. Having said that, a few caveats are in order. First, given the author’s own experience, this guide’s model article is assumed to be an empirical paper in a social science field such as political science. The insights and recommendations issued below undoubtedly apply to articles in other fields and disciplines as well, though the farther away one moves from political science, the less likely this is to be the case. In particular, articles in the natural sciences and some humanities are likely to have their own peculiarities and conventions. Readers should be mindful of this as they venture into other disciplines. Second, this post will be especially useful for reading empirical articles that broadly fall under the so-called positivist tradition. This point of departure is not meant to endorse a particular epistemological position, it merely reflects the kinds of work that students tend to encounter in my classes.
What is an Academic Research Article?
An academic research article is an efficient way for scientists to communicate their findings to other scientists. Unlike a book, which will typically cover a wide topic from a variety of angles, an article only seeks to make one or two big points. This means that an article is relatively short, perhaps between twenty and thirty pages. In the decisive majority of cases, research articles can be seen as contributions to an ongoing conversation within the scientific community. Put another way, articles are not written in isolation from existing work. Quite to the contrary, they attempt to react to what has already been said in a given field and they try to move the conversation forward, adjust it, or refocus it, but very rarely can they afford to simply ignore it. One key goal in reading an article is to figure out what the larger scientific conversation is and how this particular paper contributes to it.
The Typical Structure of an Article
The key insight that we can use to our advantage is that research articles are meant to clearly communicate some information, not to obscure it. In fact, a good research article should read like a bad movie. In a good movie, the audience is drawn in by some mysterious plot that slowly and unexpectedly develops and often includes major twists just before the end. We walk away from the movie praising its twists and turns and maybe even argue with our friends about what the ending really meant. A research article seeks to do the exact opposite. It seeks to help us understand what exactly the scientists were trying to do, how they did it, and what they found. There should be no twists or surprises, clear and efficient communication is key. This may mean that most of us will continue to watch movies on Saturday nights instead of reading research, but it also offers an invaluable starting point: research articles have a predictable structure that we can use as we read. The major parts of an article are:
The Abstract The abstract is the first paragraph you should find just below the title of an article. It should be a few hundred words long at most and it should summarize the entire article. In particular, you should be able to identify the main research topic/question as well as the findings. If you only had two minutes to read an article, you would start with the abstract. If you were looking through several articles deciding which ones you should read more carefully, you would again read the abstracts first.
The Introduction The introduction is an expanded version of the abstract, stretching perhaps several pages. It, too, begins by identifying the main research topic or question and situating it within the broader scientific conversation. This means that it provides some justification for why the topic is important and why we should continue reading. The introduction will also ordinarily briefly outline the methods that will be used and it summarizes the main findings. This is where the bad movie metaphor should be most obvious. The introduction essentially tries to give us all the main points so that there are no major surprises and twists later in the article. Some introductions will offer a brief “roadmap” for the rest of the article, others dive right in.
Literature Review and Theory The section that comes after the introduction will vary depending on the specific article we are reading but there are a few things to expect. First, the article is likely to further situate itself in the existing scientific work. It will do so by reviewing other literature in the field but only to the extent that this review actually helps the article to explain why another scientific paper is needed and why existing work has gaps and holes in it. In other words, we can expect something like: this is what others have done - this is what remains to be done/can be done better - this is what we do and how we fit in.
The literature review will then transition to more explicit statements of the research article’s theory, i.e. a system of ideas that is intended to explain something. At this point, the theory is not yet tested with actual data, it is merely posited and justified with logical assertions. For example, I could have a theory that says that as people get more educated, they read more books. This (rather shallow) theory is based on some logical supposition such as that education shifts our preferences towards reading. Note however, that the theory is not yet tested with real data, that only comes later.
The Methods Once there is a theory and some expectations/hypotheses in place, an article can then articulate how it will go about evaluating evidence to decide whether the theory is supported or not. In the previous education and reading example, the methods section could suggest that we should do a survey and ask people for their education and reading habits. This would generate data that could then be used to evaluate the theory. The methods section is often quite technical and one should not get discouraged by not understanding it very well. That is entirely expected.
Results The results section reports the findings that stem from the application of the methodology. Were the findings in accordance with theory and hypotheses? In the education and reading example, we might report the percentages of more or less educated people and how much they read.
Discussion/Conclusion The final section of a research article will typically be called “discussion” or “conclusion.” Once we have reviewed the results of a study, the question becomes: so what? In other words, the article should reflect on what its findings mean for the broader scientific conversation to which it seeks to speak. In addition, the conclusion will often make suggestions for future research and mention any shortcomings of the concluded study and how things could be done better in the future.
How to Read an Article Now that I have listed the main components of a research article and explained what each part does, how should one go about reading a scientific paper? The answer will depend on what it is that we want to get out of an article and how much time we have to get it. The key thing to remember is that one does not have to, and in fact often should not, read the article from beginning to end without jumping around. While I do recommend reading entire articles from time to time to get a feel for how the entire piece comes together, the non-expert reader will often focus on specific sections while skimming others.
Depending on one’s reading speed, reading an entire article might take about three hours or more. If we only have a few minutes and our goal is to get the very gist of the article, I would suggest only reading the abstract, albeit very carefully. This will not allow us to critically engage the piece, but it will give us an idea of what the paper is about. If we have thirty to forty-five minutes and desire a deeper understanding of the article, it is best to read the introduction and then the conclusion. Reading the introduction first will give us an expanded version of the abstract and the conclusion will likely quickly summarize the results and reflect on the major takeaway points. If we have still more time, we will want to read the theory section carefully because that is where the authors explain their ideas in greater depth. If we are reading to identify more sources to read or if we want to get a deeper understanding of related literature, the literature review will be a rewarding section. Most non-expert readers will only read the methods section last because it may be the hardest part to understand. Naturally, different readers will develop their own habits and these recommendations should be taken as suggestions based on the author’s own experience.
What to Look for in an Article As I read the various sections of a research article, I tend to look for a few key pieces of information that are useful as I write my notes and prepare for discussing the paper with others. Many of these can be located in the introduction and conclusion. First, I am looking for the broader research question. Many articles will in fact include a sentence with a question mark at the end that indicates what the question is, for example: what are the social consequences of civil war? Even if there is no question mark, almost every article’s topic can be phrased as a question. If there was no question, there would be no gap in knowledge and presumably no need to write an article.
Next I look for the argument, i.e. the suggested answer to the research question that the author(s) come up with. An argument is not yet supported by empirical evidence, it is just an educated, logical guess as to what the answer might be.
Subsequently, I try to identify the evidence that the author(s) propose to use to evaluate the merit of their argument. Are they using data from interviews? Are they working with data from the census or did they field their own survey? At this stage I am not yet thinking about what the data says, just what the nature of the data is.
Finally, I try to find a clear statement of what the results are. The results may sometimes seem similar to the argument (especially if the argument turns out to be correct) but the difference is that the results should be derived from the evidence.
How to Read with a Critical Eye The key concept at the core of the scientific enterprise is that we can objectively evaluate whether or not someone’s argument is correct based on evidence and we do not have to take their word for it. In other words, our confidence in a research finding depends not on who produced it but how it was produced. While the goal of most “casual” readers will be to understand what an article is trying to do and how it does it, it is important to adopt a critical attitude. In other words, it is important to always question what we read and ask ourselves whether it sounds reasonable.
In particular, I recommend focusing on two aspects of critical reading. First, it is important to look for the assumptions in an argument, whether they are explicitly mentioned or not. An assumption is something that the author(s) assume but do not test. For example, an article might assume that people are self-interested. A critical reader should ask whether this assumption is reasonable or whether there are conditions under which it would not hold. Sometimes the most important assumptions are those that are not explicitly mentioned, i.e. those that “go without saying.”
Finally, it is useful to think of alternative explanations. Could it be possible that an article’s evidence is consistent with an alternative argument? For example, we notice that there are more weddings in the summer and we might want to immediately conclude that this is because summer weather is nice and people want to have nice weather for their wedding. Alternatively however, we could argue that more people have vacations in the summer and couples know that more guests will be able to travel. Here the evidence is the same (more weddings in the summer months) but the argument is quite different. This is why evidence rarely speaks for itself, it nearly always needs an accompanying theory that explains its meaning.