By Ana Todorović, Oxford University
In September 2017, we started the school year at Oxford with a day of talks on robust research practices. Originally envisioned as a satellite meeting of the Autumn School in Cognitive Neuroscience, it ended up spilling over into the Autumn School as well, which meant that incoming masters students got their official welcome to the programme in the form of four lectures on scientific reproducibility.
The Oxford Reproducibility School was spurred into action by Kia Nobre, head of the Experimental Psychology department, and was organized by Dorothy Bishop, Ana Todorovic, Caroline Nettekoven and Verena Heise. And although their primary areas are psychology and cognitive neuroscience, the Oxford Reproducibility School was aimed at discussing problems in empirical science in general, as well as best research practices.
We had talks that outlined the root causes of the reproducibility crisis. Talks that discussed novel statistical approaches. Talks that covered combined academic and industry practices in pharmaceutical research. Talks about efficient computing, shared analysis pipelines, data storage and ethical practices when uploading that brain scan to an online repository. Talks that covered teaching undergraduates the right way instead of having them unlearn what they first encountered in their statistics courses. Talks about preregistration, and conversely talks about exploratory research. We ended it all with a talk on why working reproducibly is good for you personally, and not only for the wider scientific community.
The talks are now available as Oxford Podcasts, and have all been blogged about as well.