When writing multiple papers from the same dataset, how do you handle the methods section?

Question

I am planning on publishing two distinct papers based on the same dataset – one is looking at policy implications and the other is a more traditional academic article. The methods sections for both papers are largely the same (minus the statistical analyses, which are different) but the survey design, implementation, etc. are, of course, the same.

Can I use the exact same language for the methods section in both papers or does that fall under self-plagiarism and/or copyright issues? It seems crazy to figure out a completely different way of wording something when obviously it’s the same study, but I’m not actually sure what to do.

Answer 1

I intensely dislike the notion that reusing your own text across methods sections could be considered plagiarism. This is one of the stupidest bits of self-policing we've subjected ourselves to in years. There are many reasons for my objection, including:

• The first time I write something, I write it as well as I can. Writing it differently just to avoid writing it the same will make it worse. Why should I do that?
• In describing methods, precision is of utmost importance. If the methods used in two studies are the same, the descriptions should be the same to highlight this point.
• It's a waste of my time. If I write code that calls a subroutine to do a specific job, I don't code up multiple different versions of the subroutine for each call. Why should writing papers be any different?

All of that said, overreaching self-plagiarism norms are fact of contemporary academic science. If it were me, I'd duplicate the text, citing the other paper and including a note specifying that the text was duplicated. Even that may be riskier than is optimal, though.

Answer 2

There is a divide in how to approach this question based on the type of methods section, as there are two almost entirely distinct forms of methods sections in scientific publications.

• "Prose" methods: This type of methods description is part of the main flow of text, and presents the methods in an explanatory fashion. Such a methods description is expected to be either unique or else credited as being reproduced/adapted from another publication. An example of such a paper (in particular Section VI A, B).
• "Cookbook" methods: This type of methods section comes at the end of a paper, after its discussion and conclusions, and is effectively a compressed "appendix" with no narrative exposition, just an extremely terse summation of particulars. Such a methods section is not generally subject to concerns regarding self-plagiarism (and arguable not even copyrightable!) because they are essentially just recipes: there are only so many ways to say "Cells were cultured in LB medium at 37C." An example paper with such a methods section.

For your particular case, then, if you expect both papers to be under review at the same time, then you should already be making sure that the reviewers know this is the case. The papers should likely be referencing one another as "under review", the cover letter should make the existence of the other paper clear, and you should likely submit it as review-only supplementary information. If you are creating "cookbook" methods, that's enough to deal with self-plagiarism. If you are creating "prose" methods, then you should explicitly note at the beginning of the section that it is shared between the two papers, thereby making an explicit declaration of shared/adapted text and defusing concerns of self-plagiarism.

Finally, if one paper ends up significantly preceding the other, such that they are not effectively co-published, then you can simply remove its references to the other publication and change the references in the second to refer to prior rather than simultaneous publication.

Answer 3

You might want to consider having the full detailed description of the methods in the more traditional academic paper and a streamlined version (referencing the other paper) in the policy paper.

The papers have different audiences who are interested in different things. You need the detailed methods to support your results (academic paper). Having established that the results are credible, the other paper investigates different aspects.

It's a little tricky to handle during simultaneous review - but you can supply the (anonymised version) academic paper to the policy journal as supplementary information for reviewer information. Discuss this with the editor of the policy paper journal in advance.