Multiple Comparisons and the False Discovery Rate

Multiple comparisons occur when we test multiple hypotheses, whether in a series of multi-group pairwise comparisons or a series of hypothesis tests. While there is a rich and vibrant discussion as to when and where it is most appropriate to address multiple comparisons and preserving our type I error rate, we focus on introducing the general concept and some potential post-hoc solutions (more of which will be covered in our ANOVA-related lecture). The false discovery rate is another approach to control for multiple comparisons and is introduced in greater detail with examples by "hand" and in R. A video for the Biostatistical Methods I (BIOS 6611) course in the Department of Biostatistics and Informatics at the University of Colorado-Anschutz Medical Campus taught by Dr. Alex Kaizer. Slides and additional material available at https://www.alexkaizer.com/bios_6611. Table of Contents: 00:00 - Intro Song 00:18 - Welcome 00:37 - Multiple Comparisons 04:33 - Some Post-Hoc Comparison Methods 05:54 - False Discovery Rate 07:32 - FDR Algorithm 08:33 - FDR Example 11:41 - Multiple Corrections in R and SAS 12:41 - The FDR Example in R