As a mentor, the one thing I struggle the most with is communicating a good method for writing a scientific paper. It’s not that I don’t have one. It’s just that it’s completely contrary to the way we teach high school and college students to write. Many students write linearly – beginning at the beginning and ending at the end. Problem is, if you write a scientific paper that way, by the time you get to the end your story may have changed along the way. That’s very hard for a reader to follow.

So, a couple of years ago a senior colleague taught me how they story board their papers. They start in the middle of the papers at the results and write from the middle out. I’m working on a paper today and it made me a little wistful for my interactions with him. A loyal reader of the old blog was kind enough to send the post I wrote about his writing methods for re-posting here.

I still use the exactly same method, although I find that many people are reluctant to post their figures to the wall or a board and stare at them. Last night, Strange I stood in my office staring at a figure for about an hour. It made me realize that if someone needs an hour to understand what I’m getting at, it’s not a good figure. It doesn’t matter if I think it’s clear. It’s not if it takes someone that long to figure it out. Back to the drawing board.

So, without further adieu, here is a slightly edited version of the original post on the I-method of scientific paper writing…

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

1. I try to remember the original question(s) (ie, hypotheses) that I started with and write them on my board.
2. I make the figures or tables with the data that answer those questions. I take the figures and tables to the board under the question/hypothesis.
3. I ask myself what follow-up questions I had, write those on the board, make the appropriate figures and tape them to the board.
4. I tape the outcomes of my statistical analyses to the figures.
5. I make people come and look at my board. When I get them there, I let them look at my figures a while and then I try to tell them the story of what I’ve done, using my figures and tables. I reorder the figures and tables based on their feedback and how I find myself telling the story. I write the questions they as on the figures and revise. I make people keep coming back to my board [now it’s a wall in my office] until they say, “Huh. That’s a pretty good story.” That’s the most crucial step of the process – getting your story in the right order so that it makes sense when you tell it.
6. I write the results.
7. I write the methods so that it parallels the order of the results.
8. I ask myself if we have anything unexpected, how we’ve changed what we know, or what limitations we have. I write those on the figures and use that the write the discussion.
9. I go back and write the introduction based on the story that came out of the results. I am a big believer in using the phrase “We hypothesized that..” so that there is no question about what we were trying to address or whether the experiments were appropriate. This hypothesis might be different than the one that drove the initial experiments because science is not always a linear process. Don’t be a slave to the original hypothesis if you learned things after the fact that made you change gears. That creates a tortuous paper that no one should subject a reviewer to. Also, include a general statement about the approach and if its more than 1.5 pages long, it’s too damned long.
10. Add all the other stuff
11. This is the second most critical step. I give that paper to anyone that will read it and provide feedback. I will give it to my neighbor’s dog to see if he craps on it. It is better to get criticism from the people you know than the people you don’t. Anyone who can be convinced to read it gets a copy. I also pay it forward. I will read anything anyone gives me.
12. I submit it.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

After reading this wisdom for the ages you might be thinking to yourself, “How is this any different than what we do at lab meeting? We show figures there all the time. Trust me, it is. At lab meeting you flip through figures that everyone squints at. You’re trying to listen and interpret the figure at the same time and you have someone’s voice to guide you.  There is something different about looking at a figure on paper, with time to look at it and meditate over it, without someone gabbing at you about what it is supposed to mean. You see things about formatting and presentation that you don’t see on a slide.