The structure of an academic paper

The structure of an academic paper

Title; Authors; Abstract; Intro; Theory; Data and Methods; Results; Discussion; Conclusion.

With our project set up on Overleaf, we can now focus on the structure of an academic paper. It is a nested parallel structure with nine broad sections.

Title; Authors; Abstract; Intro; Theory; Data and Methods; Results; Discussion; Conclusion.

Title

First, the title. The title should convey what the paper is about. Often, many authors try to be clever with their titles by using fancy words, quotes, or other gimmicks to make themselves appear clever. However, in the Internet age, the title’s sole purpose is to succinctly and distinctively communicate to search engines what the paper is about so that people can find it quickly and be encouraged to read it further. Your title must provide transparent information about how your paper will help readers and should pique their interest without using clickbait. This requires understanding your audience and knowing which keywords will interest them. Let me give you an example of a poorly executed title.

One of my best papers is published in the American Sociological Review, where I aim to comprehend the influence of peers on student learning and their long-term academic outcomes in India. To achieve this, I collected data on the students’ social networks over four years, leveraging the randomization of roommates at Indian University to obtain exogenous variation in peer quality (inspired by Sacerdote, 2001). The paper reveals that causal peer effects are mainly driven by people you study with rather than your friends. I also found that if students are unlucky enough to have bad peers in their first year, they drop them in the future, and their low-performance effect vanishes over time. Conversely, if they make good friends, they can maintain high performance by continuing to study with them. Simple.

But here is the title:

“The Mechanics of Social Capital and Academic Performance at an Indian College”

What the heck are “the mechanics of social capital?” Are the results only relevant to people who care about “Indian colleges”? The title doesn’t explain why (a) you should care about the paper and (b) what you will learn. I might write a different title for this paper if I were writing it again today. Perhaps, “The long-term effects of peers, friends, and study partners on academic performance.” The answer isn’t easy, and thinking about how to write this title effectively may be helpful.

You can also take a data-driven approach to this task. For example, here is a histogram displaying titles from the Strategic Management Journal and the average citations of articles that employ these keywords in their titles. Although this graph conflates the information value of the title with the demand for this type of research, it may be a helpful exercise to consider which words effectively communicate the message of your paper.

Here are a few titles that might have worked better.

Beyond Friendship: A Study on the Various Types of Peers and their Effects on Academic Performance
The Mechanisms Driving Peer Effects on Academic Performance: Evidence from Randomized Roommates

Be experimental and try out different title options. Test them on a friend, colleague, or peer to see which communicates your big idea most clearly.

Abstract

Next, we have the abstract, which serves as a microcosm of your paper. This may be an excellent opportunity to discuss how you approach structuring your paper. I like to think of a paper as an hourglass.

Motivation: The beginning of your paper should be broad, serving as the motivation behind it. What big question are you trying to answer, and why should anyone care?

Problem: The next section is the problem statement, which should be more focused. Explain the specific problem that your paper aims to solve within the broader motivation you presented. There are different types of problems in research, and it’s essential to choose a fundamental problem, not just an intellectual game. As someone who has read potentially thousands of research papers, I have come to appreciate this nuance over the years. So, what specific problem does your research address?

Solution: Once you have described the problem, articulate your solution. What are you accomplishing in this paper? What new insights, tools, or data combinations are you introducing that will aid in tackling this challenge?

Your findings: After presenting your solution, share your findings. What did you discover? How did your new idea, approach, or method perform? Did you encounter any surprises or unexpected results? Make sure to be concrete.

Implications: Explain the implications of your findings for our understanding of the problem and how they should shape our perspective on the world. In other words, how does your research contribute to the broader conversation on this topic?

That’s it. On average, an abstract should consist of five concise and focused sentences that convey your paper’s key points. Each sentence should make a tight point, with some being more complex than others. The goal is to be brief and precise.

Okay, this is your abstract.

Once you have the abstract down, your next step is to write down the structure of the paper. What is nice is that there is some parallel structure to how one would write a paper.

Let’s talk about the overall structure of the paper.

Introduction

The introduction fulfills the same purpose as the abstract but in greater detail. It divides into five paragraphs, corresponding to the five sentences in the abstract.

The first paragraph should provide the motivation or hook for the reader. Why should they care about your paper?
The second paragraph should lay out the problem in more detail. For example, why do most startups fail? Why do many low-income entrepreneurs choose underperforming businesses? It’s essential to provide specific details and background information to give readers a clear understanding of the problem.
The third paragraph should describe your solution. It should detail how you built a theory about X and Y, proposed the Z method to resolve the uncertainty, and brought this approach to your paper.
The fourth paragraph should summarize your findings, including significant effects or results. It’s essential to include any relevant details, such as percentage changes and any scope conditions or limitations of the study. Please also describe any long-term effects your research has uncovered.
Finally, the fifth paragraph explains the implications of your work, both for the problem, you are solving and the literature at large.

Notice that the structure of a research paper resembles an hourglass. You begin broadly with a general motivation that captures readers’ interest, then narrow down to the specific problem or puzzle you are addressing, followed by the details of your solution and findings. Finally, you go broad again and discuss what your paper teaches us and who can benefit from this knowledge.

Now that your introduction is complete, the other parts of the paper follow this structure. The sections of your paper are the introduction, theoretical framework/literature review, methods and data, results, and conclusion and discussion.

The introduction begins with a broad overview of the paper, summarizing its contents. It should provide enough information for the reader to understand the topic and findings. From there, the paper narrows down to the problem or puzzle being addressed, followed by the proposed solution and findings. The introduction is the most comprehensive section of the paper, and readers may use it to decide whether to delve deeper into the theory and results sections.

Theoretical Framework

After the introduction, you will have the theoretical framework. This section can be challenging to write, especially for strategy and organization theory audiences, as there is a lot of variation in taste. Some authors have overly complex language and convoluted models, while others stick to a simple literature review. However, the most effective approach is to present a clear and concise theory that directly addresses the problem statement you introduced in the previous section. This theory should be grounded in the relevant literature and provide a solid foundation for the hypotheses or predictions you will later test with your data. Ultimately, the theory section should guide understanding the relationships between your variables and how they relate to your research question.

The purpose of the theory section is to guide the author in their empirical analysis. It lays out the logic behind why the empirical test is necessary and what the author expects to find in the data.

Every paper is trying to convince you of some fact.

In strategy and management research (often in other scientific fields as well), the facts often consist of a multi-part statement of the following form.

The Research Idea Canvas

THEN

EXCEPT/ESPECIALLY WHEN

BECAUSE

Examples:

IF a firm has a diverse workforce,

THEN it will have higher innovation and creativity.

EXCEPT WHEN the firm lacks inclusive practices that allow for the full utilization of diverse perspectives.

BECAUSE diversity brings in a variety of perspectives and experiences, leading to more innovative and creative solutions.

Here is another one:

IF a person receives feedback on their performance,

THEN they will improve in that area.

EXCEPT WHEN the feedback is perceived as unfair or irrelevant.

BECAUSE feedback provides information on how to improve and correct mistakes.

The theory section aims to establish the logical reasoning behind the empirical analysis supporting the above facts, such as why getting advice from peers or conducting AB testing could improve startup performance. Its purpose is to persuade readers to expect specific outcomes based on the author’s arguments and evidence while addressing potential counterarguments. This section can take different forms, including a literature review, a theoretical model using mathematics, or formal logic. Ultimately, it aims to convince readers of specific facts and the logic used to support them. The author must use supporting evidence at a conceptual level to make a convincing argument.

Classical rhetorical structure:

Exordium – The introduction, opening, or hook.
Narratio – The context or background of the topic.
Proposito and Partitio – The claim/stance and the argument.
Confirmatio and/or Refutatio – positive proofs and negative proofs of support.
Peroratio – The conclusion and call to action.

A helpful approach to conceptualize the theory section is through the five-part classical rhetorical structure. As mentioned above, these five parts comprise the abstract, introduction, and theory sections.

Like the introduction, the theory section begins with a setup – a null model of the world that most people should believe. Start by stating the importance of the problem and providing a deeper introduction to the topic by citing relevant literature and summarizing what it says. Then, provide details about your topic – what you are studying – and where the gap lies in the literature.

Next, structure your theory section as a set of claims that you are trying to get people to believe. These claims are nonobvious in a sense. A nonobvious claim is not necessarily outlandish but might not be something people have thought about deeply. Suppose one were to think about this claim solely from basic logical facts or truths we already know. In that case, one could come to a different conclusion. For instance, in the peer effects literature, it is not evident that peers will increase students’ academic performance. While some people would think this is obvious, others (perhaps economists) might believe all we observe with “peer effects” is the selection on unobservables.

However, there are other reasons this is likely to be true.

After making a claim in the theory section, your job is to provide a set of arguments that convince the reader that your claim is valid. The best way to do this is by getting them to believe in minor claims that have either been confirmed empirically or are apparent and then linking them to the truth of your claim. For example, the claim that AB testing leads to better performance was initially contentious, and we had to provide convincing arguments to the reviewers.

We needed to break down the problem by asking ourselves: why did we believe that AB testing led to improved startup performance, and what were the supporting and countervailing arguments for this claim?

If you look at the paper, and see that we made several arguments:

We began by setting up the problem: a key challenge faced by startups is uncertainty about the consequences of their decisions regarding what products to launch, what markets to enter, and how these products should look and feel.
We introduced the idea that experimentation, in general, is what leads to better startup performance.
We explained that A/B testing is a specific form of experimentation made possible by digitization.
We provided several reasons why A/B testing should lead to better performance: it reduces the cost of testing ideas, helps quantify the impact of bad ideas, and fosters organizational learning.
We connected our argument to the widely accepted idea that organizations better at learning perform better.

Now that you have presented the “for” arguments, it’s important to consider setting up compelling counterarguments to your claim. This is essential for any claim in the paper because it increases the stakes. If the opposite of your claim is compelling or might be true, your empirical test will matter more in providing evidence that may shift people’s beliefs towards your claim versus the counterclaim. If, for instance, everyone already believes your claim, then your paper may need to be more helpful in changing their minds. You are here to tell people something they didn’t know already, and presenting counterarguments helps to make your case more convincing.

As a result, you should spend some time on the counterarguments.

Here in the AB testing paper, we make a few arguments that may lead us to believe that testing might not increase performance. That is, we are setting up a counterfactual.

Firms may rely on gut instincts instead of experimenting, despite having experience and tacit knowledge that can lead to better decisions.
Prior research has shown that entrepreneurs need to be more confident and may think too highly of their ideas.
Entrepreneurs may run uncontrolled experiments, such as tweaking and learning by doing, which is a good approach, and may question the need for formal experimentation.

In addition to the primary claim, you should consider two other types of claims that would be beneficial to include in your paper. Firstly, the primary claim is typically only sometimes valid in all situations. Thus, it is essential to identify the conditions under which your claim holds. For instance, does A/B testing only improve performance in firms with an organizational structure that can effectively use the information from randomized controlled trials? Are there differences in the effectiveness of A/B testing between smaller and larger firms with different organizational structures?

A second set of claims extends the original claim: can AB testing also increase webpage views? Could it affect the number of new products startups release? Would some startups shut down when they learn they cannot increase their performance with the idea they are pursuing?

The tool I use to write the theory section is a simple outline, which follows the structure I will describe in much more detail in the next section.

Methods and Data

Now that you have written the theory section, it is time to discuss your methods and data.

The methods and data section is fundamentally descriptive. Your job here is to be as transparent as possible about how you will test your claims with the most convincing data you have.

This section may vary depending on the type of data you use (observational or empirical) but usually consists of the following three parts:

The first part of the methods and data section is to describe your data or experiment. Provide as much detail as possible about the underlying data that you will use to estimate the models that will provide evidence for or against your claim. How did you construct this dataset, and what information does it include? What is unique about this data that allows you to construct a convincing empirical test?
The second part of the methods and data section describes your variables. To do this, follow standard procedures describing constructing your dependent and independent variables. One practical approach is to write each variable as a separate paragraph, starting with the variable name followed by a colon. It is also helpful to provide descriptive statistics about your variables in the write-up, giving readers a sense of what your setting looks like empirically. This is also where you typically provide a data summary table. In addition to independent and dependent variables, control variables are another type that may help reduce the likelihood of upward bias in the independent variable.
The final section you should include is your empirical strategy. In this section, you will describe the model you are estimating, how you fit the model, and what you are looking for in terms of significance, sign, and magnitude of the coefficients on the right-hand side of the model. Additionally, it would be best to discuss how this specification allows you to deal with inference problems and whether you can make a causal claim about the relationship between an independent and dependent variable.

After setting up your data, variables, and models, it’s time to move to the results section, which is usually relatively short. Most of the work in setting up the paper is already done, and now you can present your results.

Results

A well-written result section should follow a format that is familiar to the reader based on the points you have previously explained. Readers should be able to read through your results section and understand how it connects to the claims you made in your theoretical framework section.

In the AB testing paper, we made the first claim that AB testing increases performance. Therefore, the first part of the results section presents evidence using the data as specified in the methods and data section, to test this claim. This test typically corresponds to the first table in the results section. Apart from the primary test, many authors (including myself) also include additional robustness tests, such as adding fixed effects, more controls, etc., to provide further evidence for the claim.

The other subsections in the results section test the remaining claims and are usually structured in a similar manner. Each subsection presents a sub-claim along with a corresponding table and narrative that describes the evidence for or against the claim based on the data analysis.

Another section that some papers may include in the results section is the “Robustness tests” section. This section can either anticipate criticism about your methods by trying alternative specifications or respond to criticisms arising in the review process. The structure of this section is similar to the previous sections, but the empirical tests are not tests of claims but counterclaims. For instance, one counterclaim to the claim that A/B testing leads to significant improvements may be that it only leads to incremental changes. In this case, the effects may be driven at the tails, which can be observed in the code changes.

Discussion and Conclusion

The final section of your paper is the Discussion and Conclusion section.

This section typically consists of five paragraphs.

First, you restate the motivation, research question, and the main contribution of your work.
Second, you describe the main findings of your study, highlighting the results that support your claims and any unexpected or surprising findings.
Third, you provide scope conditions, discussing the contexts or conditions where your effects are strongest or weakest.
Fourth, you explain the contribution of your work to the relevant literature, identifying the gaps in the literature that your study addresses and how your findings advance the field.
Fifth, you acknowledge the limitations of your study and discuss directions for future research.
Finally, it would be best if you ended on a positive note, discussing the potential implications of your findings for theory and practice and expressing enthusiasm for future research in this area.