Our Open Experiment
When I started doing research, I followed a well-travelled path: it starts with a scientific question, followed by the design of an experiment to answer the question, writing the necessary computer code to run the experiment, then data collection and analysis, culminating* in the glory of publication in a suitable scientific journal.
I was happy to share my methodology, code and data with other members of the team, but it had not occurred to me to share it with anyone else — or everyone else.
The more I do research, the more I think that the whole process, from initial idea to final publication should be open and accessible to everyone. I don’t want to discuss the pros and cons of Open Science here (there are excellent discussions elsewhere, see this post for a recent example). But an added benefit for me is that sharing the process can give a much more honest description of how science is actually done.
We may give the impression that the process is a well-defined and objective path from hypothesis to experiment to publication. But in practice, we design our experiment so we can answer the question as well as possible given the constraints of expertise, equipment, funding and time available. We are also regular people, with lives outside science that can seep in and introduce self-doubt and stress. I am not saying it’s all terrible — on the contrary, it is mostly amazingly exciting and rewarding. But it involves emotion and compromise, which is not something we usually associate with science.
So here’s what I’m going to do. We are, at the moment, preparing to run an experiment that is very closely related to our Royal Society exhibit. Among other exhibit-related posts in this publication, I will post the thinking behind our experimental design and the choices we’re making because of various constraints. And during data collection, I will post updates and videos (well not quite during, but shortly afterwards). I will also make my code publicly available, and as much of the data as I am allowed (I need to check what the ethics application says on data sharing).
I’m excited, but also a bit scared — what if you find that I am using the wrong method, or spot errors in my code? Self-doubt and stress aside, it will make our science better, so please get in touch with your comments and ideas!
*This is often not a single step, but a near-infinite loop of rejection and re-submission, which feels more like descending to a pit than achieving glory.
