Slides and code on multiple hypothesis testing correction procedures.
This talk discusses issues relating to hypothesis tests with multiple outcomes. It also lays out three commonly used solutions:
- Dimension reduction with indexes
- Family-wise error rate corrections
- False discovery rate corrections Code is provided, focusing on computational implementations of the methods discussed in Stata.
These slides and code cover the materials discussed during a number of talks. The details of the simulations presented in the slides and discussed in the code are contained in the file multHyp.do. To replicate this code, it is sufficient to download both files contained here (the do file and the ado file), and run the do file from the folder containing the ado file. No other changes should be necessary for replication. Further information is contained in the header of multHyp.do.
This code has been tested on Stata 14, 15 and 17, though likely should work on earlier versions of Stata.
Queries can be addressed to dclarke[at]fen[dot]uchile[dot]cl