When should you stop doing science and start writing a paper?

Start writing as soon as you have at least one result.

This is a lightly edited version of a blog post I originally wrote in 2013.

In my experience, most graduate students and many postdocs think of science as a two-step process: First, you do science. You do experiments, collect data, write code, carry out statistical analyses. Then, once you’re done doing science, you sit down and write up your results for publication. Within this mental model, an important question arises: How do you know that you have done enough science? When should you stop collecting data and start writing a paper?

Photo by National Cancer Institute on Unsplash

I’ve seen plenty of graduate students struggle with getting papers out because they can’t convince themselves that they have done enough science and are ready to start writing. There’s always another experiment to be done, another analysis to be run. In my opinion, the fundamental issue here is not that these students have difficulty judging when they have done sufficient work. Instead, their mental model of science as a two-step process is flawed. Realistically, you will never have done so much science that there is nothing else to do. And at the same time, even if you think you have a complete story, you almost never do. Anybody who has written at least a few papers will have experienced the situation where, as you’re making your case for one conclusion or another, you realize that you could make a much stronger statement had you done another experiment, or at least another analysis. What happens then is that you go back to the lab, do the work, and subsequently use the additional results to write an improved manuscript draft.

Therefore, in my mind, the question “when should you stop doing science and start writing a paper” is ill-conceived. Doing science and writing are two sides of the same coin. They should happen at the same time. In my mind, there is an obvious simpler question, and it has a simple answer. The simpler question is “when should you start writing a paper?” The answer to this question, in my mind, is “as soon as you have at least one result.” As soon as you have a result, you can write it up. Make a figure, add a caption, write an explanatory paragraph or two, and put it into a document section titled “Results.” Place a brief outline of the methods used into an accompanying section titled “Methods.” As you keep doing more science, keep adding your results to the document, and see how the story unfolds. If there are obvious gaps, think about the work you need to do to fill them. Once you think that you have a reasonable story, all you have to do is add an introduction and a discussion and you’re done writing your paper. Submit.