Tutorial in Bayesian Statistics Part 1: Basics and Bayes factors

2020.06.17
Presenter: Ronald van den Berg
Part 1 reviewed frequentist hypothesis testing (based on p values) and contrasted this approach with Bayesian hypothesis testing (using Bayes factors). Basic concepts of Bayesian statistics were reviewed (posteriors, priors, etc) and several standard hypothesis tests were discussed from both the frequentist and Bayesian perspective, including correlation, t-test, and ANOVA.
Prerequisites: basic probability theory, basic frequentist statistics
Slides: www.cns.nyu.edu/malab/static/...

[@neocrom922]

0:00 Introduction: why stat. tests, hypothesis
3:54 History of Frequentist Statistics (Fisher vs. Neyman-Pearson)
6:18 Example for Fisher NHST and Critique
10:09 A P-roblem: p values in high profile criminal cases
18:15 The Bayes Factor
27:11 Taxonomy for Interpreting Bayes Factor
34:43 Comparison (side by side) NHST and Bayes
Example 1: Correlation Analysis
41:22 With Frequentist Approach
48:50 With Bayesian Statistics
58:00 Statistical Software for Bayes
Example 2: t test
1:07:28 With Frequentist Approach
1:09:12 With Bayesian Statistics
1:12:46 Take Home Points

[@NickGhale]

Time stamp god

[@ProfFeinman]

Isn’t the Bayes Factor the same as Relative Risk and, as in the presentation, it depends on the experiment, that is, not the mathematics. That is the philosophical deference between Bayes and frequentist, as you say. Oddly, some frequentist point to 2:1 on RR or odds ratio and that is usually what you need to have epidemiology entered into evidence in a court of law.