3-3-0

In science we often have to decide whether data favor one hypothesis or another. This is easy if our evidence excludes all but one hypothesis. In practice we often face a harder task: several hypotheses remain that are compatible with the data, although one may seem the more likely one. How de we evaluate which one is more likely in the light of the data, and what exactly does more likely mean here? This course examines tools designed to help us in selecting the hypothesis that receives most support from the data: Neyman-Pearson P-values, Fisherian P-values, Bayesian posterior probabilities, confidence intervals, Bayesian credible intervals, and information-based criteria like AIC and BIC. You will learn how to calculate and how to interpret P-values and true measures of evidence, and how to use them in your own research.

Prerequisite: PHY101a

Professor van Hulst