It’s Open Access Week! What are you doing to celebrate? Decorating the Open Tree on Open Eve? Having a big Open dinner with your extended family? Celebrating with quiet introspection, followed by a Skype group call with a bunch of like-minded individuals from INNGE, then turning your thoughts from that discussion over in a blog post (this is the traditional form of celebration among my people).
So, today, I’m thinking about Open science, for the people.
Every week, in lab meeting, we have a discussion on some topic, usually taking the form of reading a paper about cover crops or something,1 and it eventually turning to a discussion on the philosophical aspects of science, data interpretation, the universe, and ourselves as people.
It has been brought to my attention by my labby brethren that I use the phrase “There is a movement in the open science community…” in almost every discussion. Does this mean I’m a one trick pony? I hope not. There’s a lot of movements in the open science community, after all.2 But one thing is clear, at least in my field, there’s a sense of ‘otherness’ associated with open science. It’s, on some level, that nutty thing the idealistic postdoc (i.e., me) is always prattling on about like it was the best thing ever, on another level, it’s a set of potentially risky, and definitely time consuming practices.
There are a lot of good reasons to practice Open Science. Reproduciblity. The ability of science to meaningfully build on itself, rather than act as isolated case studies.3 To democratize access to information. To allow the public to see the details of the scientific process so they don’t see academia as a group of insular nerds. These are a good things, and I’m sold. Most scientists will agree. But. One of the things us starry-eyed idealists in the Open Science community are really bad at is acknowledging that there are a lot of disincentives to open practice. Below are some disincentives I’ve personally heard articulated when I’ve been pontificating on Open Science:
Share your data? But I’d have to change how I set it up, that’ll mess with my system/people may scoop me/the data is mine dangit/I won’t get credit if someone uses it
Share your code? Version control? Um, no, that’s a terrible idea. Besides, Git is an obtuse, practically impenetrable rabbit hole of brospeak4/Also people will scoop me
Preprints? Are you insane? The journal I want to publish in will not publish things that are previously published, I know a guy whose paper got rejected because it was based on his dissertation.
Publish in an open access journal? Aren’t those predatory?/in my field, people only publish in OA if they’re desperate/ Um, PLOS is so expensive that we can really only afford to publish there on special occasions.
Practice open science, in general? I wouldn’t know where to start/I don’t have time/I’m doing fine without it/ Be quiet Christie, you’re annoying.
With all these disincentives in play, it’s really hard for the open-curious5 to figure out how to break in. Especially if it’s presented in an all-or-nothing monolith. Especially if the open science community receives movements towards open practice with a derisive “Yeah, but why aren’t you doing more?”
We should see more of this:
This is why I advocate the “Baby steps” approach to opening science. Helping our colleagues take small, meaningful steps towards open practice that are directly beneficial to them.
For me, data sharing has been the easiest, most obvious ‘in’ to open practice, but your mileage may vary.
I’ve written about data sharing before, and why it’s important to me. I’ve found a niche for myself, helping others manage, analyse and interpret their data. It’s translated to a very productive year for me (six papers, and counting). I only had a really meaningful role in collecting data for one of these papers, so I imagine this sounds pretty self-serving. But, I’d argue that this was a very symbiotic relationship with my co-authors- by sharing data with me, we, together, were able to complete analyses faster, using different ideas and perspectives. I would dare say, in fact, that sharing data has helped both me and my colleagues be more productive.
Open science is an ongoing process that we can strive for. Despite my own advocacy, I’m not a 100% open scientist myself- it depends on the situation- whose data I’m working with, how amenable my colleagues are to the practice, if logistics/constraints allow for it. In general, I try to share data, code, presentation slides. I believe in fully scripted analyses for both reproducibility and save-myself-a-headache reasons. I have never published a preprint. My use of OA journals is sporadic at best, because I am on the job market.6
I believe Open Science is a lofty goal that has to be worked toward in bits. It’s hard to change practices when disincentives loom large and benefits aren’t tangible on an individual level. So: open science advocates, I challenge you to help your colleagues take a baby step that benefits them. To the open-curious, don’t get overwhelmed. Find a step you’re willing to take, and try it. Publish in an OA journal! Take a datacarpentry workshop! Share a dataset after you publish on it! Tell me about it. I’ll give you a high five (virtual or otherwise). Let’s celebrate Open Access week all year long.
1 cover crops: less consistent effect on arthropod abundance than you’d think.
2 Although honestly, I probably harp on open data more than anything else. I guess I’m a data pony, though I like to think of myself as a data unicorn.
3 From a paper I read back in the day, while putting together my PhD proposal- it really resonated with me: Nakagawa, S. and I. Cuthill. 2007. Effect size, confidence interval and statistical significance: a practical guide for biologists. Biological Reviews 82:591-605. It may, in a burst of irony, be paywalled. I can’t tell from my privileged Ivory Tower IP address. And want to know something else scary? I just went back and looked at my PhD Proposal. Oh, how young and naive I was then!
4 Github, you know I love you, but when you tell a n00b to fork your repo without a gentle explanation of what that means, that n00b is very likely to say “well, fork my repo, too! I’ma just going to go do a calculation in excel and send my manuscript off, mmmkay?” I was very motivated to learn, and it still took me what felt like forever to understand when to push, when to pull, and when to commit. And (unpopular confession) I still do most of these things through a GUI.
5 A term coined during the INNGE skype call yesterday.
6 My current major incentive is proving my worth as a scientist to those making hiring decisions, and my perception is that publishing in conventional, respected journals is still weighted heavily in this circle. I have often, upon receiving a rejection due to ‘fit’ or ‘impact’ been heard to mutter (or yell repeatedly) “Well, screw you guys! When I have tenure, EVERYTHING IS GOING TO PLOS ONE!” Although, I admit, it would also really stroke my ego to get a paper in Nature. After all, being interviewed by Nature was a pretty good day for me.