How to write a Methods section

Our relatively young group has nine graduate students, and a number of them are starting to write up their first manuscripts for submission. To help prepare them to be better writers and communicators in the primary medium of our trade—scientific publications—I’ve started to present some useful pointers on how to actually go about organizing and writing a manuscript to be submitted as a scientific journal article as part of our Hot Topics series. I’m sharing this material in case others find it useful! Please feel free to provide feedback or suggestions as to how we can improve these resources and suggestions!

Where should I start?

There are some great resources available for writing scientific papers:

Guidelines for writing a Methods section

Choose a structure that works for the material you want to present.

Traditionally, papers in chemistry or biophysics have a relatively rigid structure: * Introduction * Methods * Results * Conclusions

Conforming to such a rigid structure often makes it difficult to tell the scientific story. Rigidly conforming to this structure is not always necessary with modern journal formats. You can often find a structure that allows you to tell the story without having too many irrelevant details get in the way, making sure to include complete details in an appropriate location so that the work is fully reproducible by other scientists and evaluable by refrees.

Be empathic!

Provide sufficient detail for someone to reproduce your work.

You will need to provide sufficient detail for a competent researcher in the field to reproduce it.

Keep the flow clean.

Don’t clog up the main flow of the story with irrelevant details, but be complete by including what is needed somewhere (e.g., separate Detailed Methods section, Appendix, etc.).

Examples of good Methods sections

The best way to get a feel for what makes a good Methods section is to read a lot of them, and judge for yourself which structures work well in effectively communicating what was done and why. Here are a few papers we have written that we feel are good examples of different strategies for doing this:

Computational papers

Experimental papers

Writing a computational Methods section

Writing an experimental Methods section

Is there a “reproducibility crisis” in science?

Recently, a number of journals have devoted a lot of thought and discussion to the “reproducibility crisis” in science. Some of these problems have been blamed on poor practices for presenting research methodologies. The Nature Publishing Group has collected a number of perspectives and examinations of the subject here:

Can we go further to make work more reproducible?

New technology makes it easier to share the exact version of the code to perform the study and analyze the data, and potentially even isolate against upstream software issues by emulating the platform on which the work was performed: