Tips for Writing Scientific Work More Effectively

Being brilliant in scientific ideas is not enough; we must also be able to communicate those ideas clearly and persuasively, otherwise even strong insights may go unnoticed. One of the primary means of communicating science is through writing—whether in papers, proposals, or technical reports.

Done properly, writing is not easy. It is a demanding intellectual process that forces us to clarify our thinking, sharpen our arguments, and confront weaknesses in our reasoning.

Reflecting on my experience publishing papers during grad school and postdoc, as well as conversations with my advisors, I have collected the following principles to help communicate research ideas more effectively.

What I think to be a process for effective scientific writing:

1. Determine the objective of the paper—this can be the main conclusion.
   • Have an elevator pitch: understand what you’re doing, why you’re doing it, why people should care, who should care.
2. Come up with a cohesive narrative for the paper.
   • Create an outline from start to end to convey the objective of the paper.
   • Don’t be too vague with the items in the outline: put sentences instead of just words or phrases.
   • Sketch figures you want to put.
   • Iterate and put in more and more details. This WILL take several iterations.
3. Develop the outline into paragraphs.
   • Write whatever is in your mind.
   • Write like a drunk.
   • Don’t worry about grammar and sentence structure. You’ll polish it later with the help of LLM

Suggestions on the order of the sections to work on:

1. Methods: This is usually the easiest one to write. However, its content also depends on whether you know what you’re going to put in the Results.
2. Results: You may work on this section concurrently with the Methods. Make sure that the story (especially the ordering) in the Results section is narrated towards achieving the objective.
3. Discussion: Here, you take the common knowledge in the field and bring it towards your observations in the Results section. If done well, writing the Discussion section can be as hard as the introduction.
4. Introduction: Often times this is the harder part to write. So, don’t start with it. Here, you start with motivating a broader audience, then funnel it down to what you’re work is about.
5. Conclusion: I think this is easy to write if you really know the objective of your paper and the results. But, probably you want to write it towards the end of your process just to make sure you have the entire story.
6. Abstract: And this would be the last one to write, because you really need to know the entire content of your paper. However, usually writing Abstract is not as taxing as the other sections.

Suggestions on using LLM:

• ALWAYS check what LLM tells you. Don’t trust it completely.
• You may use it to organize your thoughts or overcome writer’s block.
• Provide your content as much as possible.
  • For example, you can draft paragraphs first and use LLM to polish them.
  • Use LLM on sentence-by-sentence or paragraph-by-paragraph basis.

Additional resources that I found to be useful:

• How to Write a Good Scientific Paper by Chris A. Mack