8

I regularly teach an undergraduate course in which students are expected to complete projects that involve writing code. I also supervise undergraduates in other project and thesis work that involve writing code.

I am considering requiring future students to publicly release their project code, and any other materials needed to reproduce their work, under an open source license. (This would, of course, be stated up front in the syllabus, or before I agree to supervise a thesis student.)

(Fortunately, the research platform on which my students work is very well suited for reproducible research, having been designed specifically with that goal in mind.)

On the one hand, I have this vague idea that it's to the students' benefit to practice participating in open, reproducible science.

On the other hand, students may feel reluctant to release their code (for all of the reasons documented here). Furthermore, I want to avoid things that distract too much from the main goals of the course, which don't directly involve reproducible research.

Are there any specific pedagogical arguments for or against instituting this requirement?

(Note that I'm asking about the pedagogical impact on my students, not the benefit to the broader research community of having my students do open science.)

Community
  • 1
ff524
  • 108,934
  • 49
  • 421
  • 474
  • 2
    Would not it be better to learn them to use version control (SVN, git), wikis (e.g. redmine) for documentation, bugzilla for filing bugs and something like trello for collaboration? This will also help them, besides research, in their future work in industry. – Alexandros Sep 30 '14 at 11:23
  • 1
    @Alexandros Software development practices and reproducible research practices aren't mutually exclusive, there's no need to choose. But this question is specifically about the latter. – ff524 Sep 30 '14 at 13:46
  • If you are in the US, there may be FERPA-related barriers to requiring students to publicly release their code. My university lawyers consider the mere fact that a student is enrolled in my class to be a FERPA-protected educational record. – JeffE Oct 15 '14 at 11:45
  • @JeffE I read this article which suggests ways to avoid FERPA issues in a related scenario – ff524 Oct 15 '14 at 11:51

2 Answers2

1

I think one good pedagogical reason to not require students to use open/reproducible research is that it forces them to engage in a public display of their scholarship, rather than offering them the opportunity to do so. In a learning situation, trust, risk, and power are always being negotiated. By requiring them to release it to a wider audience, you are setting some significant boundaries on the negotiation. Instead, is there a way you can set it up so they follow standard practices and documentation, but the "openness" is only with you (or their classmates)?

Also, as a feminist researcher, I always consider how my scholarship could be misused or misapplied to achieve someone's agenda. I have done work that I would not publish or want to make openly available. I am not willing to commit to public presentation of my research before I conduct it and therefore, I would not require that of my students.

user22934
  • 41
  • 2
  • The suggestion at the end of your first paragraph kind of misses the point; I'm asking about giving my students an experience with openness, not documentation or other practices not directly related to openness. And I think there's a big difference between imposing this requirement on all of a students' research, and imposing it on a single class project (that isn't really anything I would call "research"); your latter point could apply to the thesis students, not so much the class project. – ff524 Oct 15 '14 at 04:05
  • That last paragraph sounds an awful lot like intentional publication biasing, bordering on outright dishonesty. –  Jul 11 '16 at 20:12
1

I applaud your willingness to further your student's contributions to open source.

There are lots of pros:

  • Build student's reputations (ancillary to professor and institution)
  • Create value for society
  • Students will have an even greater stake in code they know will reflect on themselves, and as a result will go to greater lengths to ensure they demonstrate learning and excellence.

Con:

  • Students (and stakeholders) may not wish to publish because of fear of negative impact to reputation.

Well, masters' theses and doctoral dissertations are, in most institutions that I'm aware of, are all supposed to be bound and stored accessible to the public, even if the public never reads it.

In all likelihood, the bad stuff will get rarely read or used. The good stuff will never get enough attention, but it will get far more than the bad.

Conclusion

It's a risk worth taking, but bounce it off your department head and be certain that they agree. They may decide that since they're undergraduates, it's unfair and not proportionate to require them to publish. Or they may insist that the requirement appear ahead of time in the course catalogue description.

Aaron Hall
  • 2,369
  • 1
  • 15
  • 34