Just to illustrate it by example:
Given 3 groups, I want to examine whether there are significant differences in proportion of female sex for all pairwise comparisons after adjustment for false discovery rate (in SAS).
I know proc glm is not advised in this case since it is for continuous outcome but I assume there may be an analog in proc logistic to adjust for multiple comparisons.
Thanks for hints and helps!
PROC MULTTEST can be used to make adjustments on p-values for pairwise comparisons, including false discovery rate. I haven't used it in a situation similar to yours, but I have used to for adjusting p-values.
Here's an example where you can input the raw p-values and then MULTTEST will provide FDR-adjusted values.
https://support.sas.com/documentation/cdl/en/statug/63347/HTML/default/viewer.htm#statug_multtest_sect026.htm
- Badges
- Science topic