Writing a Research Paper
So, you're writing your first (or 100th) paper, and you're in that stage of "Where do I start?" or maybe you've started writing and it's just not coming together. There are countless approaches and guides to writing; this guide represents my approach. Hopefully it is a useful place to start.
Most research papers are best structured by answering the following five major questions (with some subquestions). This roughly corresponds to the content of the final paper. As a first step, try to clarify to yourself in writing what your answer is to these questions.
Why does it matter to the world, to humanity, to the research community? This is usually in the Introduction and focuses what you are emphasizing and de-emphasizing in the rest of the paper.
In academic writing, you always position yourself with respect to an existing literature. Having a good literature review will make it easier for you to understand the contribution. However, there is also a risk to dive deeper and deeper and never leave the endless web of links and references.
When working on a literature review, you will find yourself going back and forth between reading/searching and writing. A realistic approach is: Start with asking what general literatures you want to review, and keep it at a high level at first. Search for papers, and write down a few sentences about each paper you find. Then put them in your paper draft.
Each paragraph of the lit review is usually about one subset of the literature. It is good practice for each of these paragraphs, to end by answering the question: What is missing in that literature? This gap is what you are addressing in your work.
In some cases you will summarize related and similar works by emphasizing how what you are doing is different than what the cited reference did.
The core part of your paper. This should come out directly from the previous sections. I.e.,
In technical papers, this is the mechanism, algorithm, or framework. In studies it is the experimental design and procedure.
You should then have a clear idea of how you evaluate whether you have addressed this gap. This depends heavily on the kind of research contribution you are making.
This section evaluates the outcomes of the evaluation. Again, this is highly dependent on your particular research contribution. When writing your findings, do not simultaneously interpret or discuss them. This is best left to a separate section.
After describing the findings, outcomes, and other results, you follow up with a Discussion of the results. This also can include the impact, limitations, and possible extensions of your work. You can be more speculative in this section.
At some stage you want to write an outline of your paper. This is a crucial step to make your paper flow and make sense to your readership. In my practice, when I outline, I often start with writing a first draft of the Abstract to lay down the whole argument according to the big questions above. This Abstract will eventually be deleted and rewritten at the end of the process.
Then you should write down all of the section headings, including subsections. For each subsection, write down a topic sentence which is a placeholder for the main idea of that subsection.
You want to have one topic sentence per "future paragraph" in the paper. These sentences should not describe your to-do work ("Describe the results from the questionnaires") but instead should stand for the paragraph ("User satisfaction was rated higher for the pre-emptive treatment").
When you're done outlining, the whole paper can be read in topic sentences alone and it should flow and make sense. You will rework this outline until that is the case. Take a look at this guide for how to think about major and minor points in your outline.
- Writing Tips for Ph.D Students by John Cochrane - Agree with almost all of it, but some of it is too extreme for reviewers and editors.
- Short tips from Zachary Lipton - Don't agree with all of it, but most of it is spot on.
- Good Writing by Marc Raibert - Good tips, but a little wordy; ironic for a writing-advice column.
- How to Write a First Class Paper
- Four Steps to an Applied Micro Paper - General tips for empirical research
- Whitesides' Group: Writing a Paper
- Ruina's Report Writing Advice - Great tips for those who don't think of themselves as "writers"
- Excellent Tips on Good Math Writing