Research Paper 101

- 8 mins
research writing

I think one of the most painstaking parts of scientific research is actually writing the paper. Sometimes, this can even feel harder than conducting all the experiments. The main reason behind this, I believe, is a lack of motivation. Let’s be honest, writing can be pretty boring. Most people don’t feel nearly as enthusiastic about writing the paper as they do about writing code or testing out theories. But the truth is, this step is just as important as any other part of the research process. If you can’t express your ideas clearly (and make them sound compelling), all your hard work might just go unnoticed.

I’m still at an early stage in my research career. As of writing this article, I have two conference papers, a Q1 journal paper, and another paper in the second stage of review in a Q1 journal. This guide isn’t something I came up with on my own. It’s more like a compilation of the steps my supervisors advised me to follow. I’m mostly writing this down here for myself, so I can refer to it whenever I feel stuck. At the same time, I hope putting it online helps someone else, too.

The sections are organized according to the recommended writing sequence.

The Very Beginning

For your literature review, focus on selecting studies from reputable journals and conferences. Let’s say, you’re doing something that has already been done and published in a Q1 journal. Now, you want to do a similar thing but on a different domain. It’s natural to hesitate in citing the original work, fearing it might diminish your own contribution. However, literature review is not just about finding a research gap. It also showcases successful methodologies. In this case, if you mention the study in the Q1 journal, it gives your work a bit of credibility because a similar approach has already found success.

On the other hand, if you’re proposing a completely novel methodology and cite papers from well-known journals, it demonstrates that you are confident about your work. It signals that you’re well-versed in the existing landscape of research and are presenting a new approach with a full understanding of established practices.

Even before you start writing the paper, it’s a good idea to create a table or chart that helps you see the strengths and weaknesses of your work. In this table, you can list important research parameters in rows, and populate the columns with information from your paper and other relevant studies. These would typically be the papers you’ll mention in the literature review.

For example, if you’re working on deep learning, the rows might include: “dataset info (source, count, validity, preprocessing, etc.)”, “target/field of application (if applicable)”, “method (main points, preferably in bullet points)”, and “results (metrics like accuracy, recall, etc.)”. This table will help you in three ways:

Methodology

It’s a good idea to start with the methodology section as soon as you finish your experiments, while the entire process is still fresh in your mind. ​​Writing this part early ensures that you accurately capture what you did, step by step. It also helps you maintain clarity and precision, which are crucial for others trying to understand or replicate your work.

Writing methodology is often pretty straightforward. Here, you simply explain what you did. One piece of advice I got from a supervisor is to describe your work in such detail that even someone unfamiliar with the field can understand it. I found this approach really helpful. Writing with this mindset ensures you cover both the fundamental and novel parts of your method.

In addition to describing your process, justify the choices you made. If you came up with a few steps on your own, explain how they work and why they matter. At the same time, many parts of research rely on established methods. For instance, you might introduce a new convolutional neural network architecture as your contribution, but you’ll probably also use standard components like pooling or activation functions. In those cases, cite the original sources and explain why you chose one method over another.

Results

This section is also quite simple. Present your findings, graphs, charts, metrics, or anything that shows the outcome of your experiments. Most researchers simply report the numbers here without diving into explanations (that’s usually reserved for the discussion section). The goal is to inform the reader of the results, not necessarily to interpret them. Some people do include brief comparisons to related works, which is fine too.

Discussion

According to many researchers, this is the most important part of the paper. This is where you present your work and say, “Hey, look at what I did. Pretty interesting, right?” You interpret your results and explain why things turned out the way they did.

If your research includes multiple key aspects, it’s a good idea to discuss each one in its own paragraph. For example, if you’re comparing two methods, discuss one in a paragraph and the other in a separate paragraph.

Also, compare your work with others. Here’s where the table you created at the beginning becomes useful again. You can even include it in this section and use it to highlight the key differences between your work and existing studies.

Don’t forget to mention the limitations of your work. You might worry that this will undermine what you’ve done, but that’s not true. Acknowledging limitations shows maturity and opens the door for future research. For instance, if your study proposes a new method for identifying tuberculosis drugs, you could mention how this process might also be adapted for other diseases. That doesn’t weaken your contribution. It actually strengthens its impact by suggesting broader applications.

Introduction

This is the first section of your paper, so it’s crucial to make a good first impression. Ideally, the introduction should have four paragraphs:

If you look at papers from reputed journals, you’ll often see this same structure in their introductions.

Literature Review / Related Works

Some journals don’t require this section anymore, but I still think it’s a valuable part of a paper. This is where you talk about the other studies that inspired or relate closely to your own.

Try to include at least three papers from respected journals that are directly connected to your topic. Of course, if your work is in a very new area, you might not find that many. For each of these papers, describe what the authors did, how they did it, what they found, and where their work falls short. Be respectful when critiquing and only contradict with facts.

For instance, if another study uses 3D images for a classification task, you could mention that 3D image processing requires high computation. Additionally, you can mention 7-8 more papers in brief, just a line or two about each. These papers can be related to the overall topic (air pollution, for example) rather than being highly relevant to your work. This isn’t strictly necessary, but it adds a bit of depth to your presentation. In the last paragraph of this section, you can point out the research gaps and mention how the proposed method differs from others.

Conclusion

Here, you summarize your paper and suggest possible directions for future work. You can highlight key features of your study, but avoid repeating things word-for-word, especially result values. Keep it concise and thoughtful.

Abstract

The abstract is another summary, but this one packs everything into a single paragraph. A lot of people will judge your work based on the abstract alone, so it needs to hit all the right notes.

Think of it as a compressed version of your introduction. Start by describing the problem and the gaps in current solutions. Then, introduce your approach and mention the key results. With the information you’ve already gathered, especially from the earlier table, you should be able to write a strong and informative abstract.

After finishing writing these individual sections, it’s time to combine them. While doing this, read the full paper and remove any information that you feel is unnecessarily mentioned more than once. The usual flow of these sections is to place abstract, introduction, literature review, methodology, results, discussion, and conclusion. However, some journals often want a different order. So, always check the “guide for authors” section before starting a submission.

Hope this guide helps you craft a killer research paper! Now go out there and make your mark.

This article is also live on wordpress.