Scientists value transparency and reproducibility, but are rewarded for highlighting the novelty of unexpected findings. This is one reason why published research findings are hard to reproduce. See, for example, the recent work done by us and the scientists involved in the Open Science Collaboration on Estimating the Reproducibility of Psychological Science (http://ift.tt/1mbUvcD). When scientists preregister their research, they are making key decisions without being biased by the data they collect, which makes standard statistical tests more effective. Though preregistration is required by law for clinical research involving human medical studies, it is not widely practiced by most scientists. We at the Center for Open Science have $1,000,000 to hand out as prizes for researchers who publish the results of their preregistered research. See https://cos.io/prereg
I'll be back at 12 pm ET (9 am PT, 5 pm UTC) to answer your questions, Ask me anything!
No comments:
Post a Comment