“Dealing with Selection Bias”
Hosted by the Cognitive, Computational and Systems Neuroscience Pathway (CCSN)
The workshop is open to anyone in the Wash U Neuroscience Community.
Abstract: Selection bias occurs when the way a statistical sample is obtained prevents the sample form accurately represent the population about which one wishes to draw inferences. Straightforward as the issue may seem, selection bias is among the most pernicious perils of statistical inference. In this lecture I will discuss some of the many ways that selection bias and related phenomena such as right censoring and the Will Rogers effect can arise in research. I’ll consider what we can do about it as consumers of scientific research, i.e., as readers, and as producers of research, i.e., as authors.
Bio: Dr. Bergstrom is a professor of biology at the University of Washington. His research interests include behavior, evolution and computational biology. He also has an interest in metascience and how norms and institutions shape scientific knowledge. He is coauthor of the book Calling Bullshit: The Art of Skepticism in a Data-Driven World, which takes aim at oversold arguments and provides guidance on how to spot them and counter them.
For inquiries contact Dennis Barbour.