CCSN Research Methodology Workshop: Kevin J. Mitchell
Monday, October 16, 2023 2:30 PM to 4 PM
About this Event
Presenting on "Defense against the Dark Arts - or how not to fool yourself with stats".
Kevin J. Mitchell, PhD, associate professor of genetics and neuroscience at Trinity College Dublin will speak on Monday, October 16, 2023, at 2:30 pm. This seminar is virtual and registration is required.
Abstract: The statistical methods used in many areas of biological research were invented to help us figure out what findings we should be impressed with. Many of the experiments and analyses we do involve noisy data, with lots of variation that is unrelated to the hypotheses we are exploring or testing. Our statistical tools are supposed to help distinguish spurious findings from ones that look more likely to reflect real effects and that deserve further study. In short, these tools are intended to stop us fooling ourselves. When misapplied, however, they do exactly the opposite. In recent years, we have become more aware of the pitfalls of the naïve application of statistical methods and the problems of irreproducible research across many fields. I will discuss painful lessons learned from psychology, genetics, and neuroscience that illustrate these problems and a range of solutions that can promote much more robust and replicable research.
Event Details
See Who Is Interested
+ 1 People interested in event
User Activity
No recent activity
About this Event
Presenting on "Defense against the Dark Arts - or how not to fool yourself with stats".
Kevin J. Mitchell, PhD, associate professor of genetics and neuroscience at Trinity College Dublin will speak on Monday, October 16, 2023, at 2:30 pm. This seminar is virtual and registration is required.
Abstract: The statistical methods used in many areas of biological research were invented to help us figure out what findings we should be impressed with. Many of the experiments and analyses we do involve noisy data, with lots of variation that is unrelated to the hypotheses we are exploring or testing. Our statistical tools are supposed to help distinguish spurious findings from ones that look more likely to reflect real effects and that deserve further study. In short, these tools are intended to stop us fooling ourselves. When misapplied, however, they do exactly the opposite. In recent years, we have become more aware of the pitfalls of the naïve application of statistical methods and the problems of irreproducible research across many fields. I will discuss painful lessons learned from psychology, genetics, and neuroscience that illustrate these problems and a range of solutions that can promote much more robust and replicable research.