How to involve new collaborators in open science?

+6 votes
39 views
asked Aug 16, 2015 in Open Science by Thomas (915 points)

I frequently find myself working with collaborators that are new to open science and may not use the same tools that I use (R, knitr, git, etc.). How can/should I negotiate with collaborators to ensure that our project remains open even when they don't have experience working openly?



This post has been migrated from the Open Science private beta at StackExchange (A51.SE)

1 Answer

+3 votes
answered Aug 17, 2015 by Carlisle Rainey (70 points)

I think this requires answering two questions:

  1. Is closed science better than no science at all?
  2. How important are the various components of open science?

I'd answer yes to the first question. If you answer no, then your conclusion will be different.

To the second question, I think there is a clear prioritization. The most important component of open science in my opinion is a reproducible workflow. There are at least three components of a reproducible workflow:

  1. Keep the original, raw data and a source for it.
  2. Use scripts to manipulate the raw data.
  3. Use scripts to analyze the data.

In sum, a reproducible workflow allows the researchers to easily go from the original, raw data to the final analysis.

I try to convince coauthors that a reproducible workflow is important, not because of my open science values, but because of our own self-interest. A reproducible workflow is a much more efficient workflow, because we can easily re-create what we've done when it comes time to handle an invitation to revise and resubmit. We can easily change parts of our analysis. We can easily put together our replication files when required by the journal. We can easily see we've done when it comes time to describe it in the paper.

Other tools might not make sense for certain coauthors. For example, I never try to get Word users to switch to LaTeX or markdown. It simply isn't in their or my interest for us to spend large amounts of time learning new tools that offer only small benefits to us.

Public version control is another example. It is an approach to science that I believe is important. But I never force this on coauthors. I do use it on my end, and I version-control their changes as best I can.

The key to a successful open science movement is to (1) clearly articulate how open science benefits individual researchers and (2) change institutions to handle the disincentives for open science (i.e., preregistration).

When working with coauthors, I think the best approach is to think about the best way to produce a great paper. Where open science helps achieve that goal, make the case for open science.



This post has been migrated from the Open Science private beta at StackExchange (A51.SE)

Welcome to Open Science Q&A, where you can ask questions and receive answers from other members of the community.

If you participated in the Open Science beta at StackExchange, please reclaim your user account now – it's already here!

e-mail the webmaster

...