Multiple hypothesis testing commands in STATA (Wyoung and rwolf)

[Lanna Ng]

Hi,

I am correcting Multiple Hypothesis testing using wyoung and rwolf commands in STATA. Each command reports the model p-value, which I understand as the individual p-value of each model before applying the correction. In this sense, the model p-values should be the same in both methods. However, they are very different and different from the results when I run each dependent variable independently.

Would you happen to know what is going on here? Why is it like that?

Thank you.

Lanna

[Julian Reif]

It's hard to comment without a specific example. In wyoung, the value of p corresponds to the unadjusted p-value. In the example below, you can see that is equal to the p-value for displacement when running each regression independently.

sysuse auto, clear
wyoung mpg headroom turn, cmd(regress OUTCOMEVAR displacement length) familyp(displacement) bootstraps(100) seed(20)

regress mpg displacement length
regress headroom displacement length
regress turn displacement length

[Maxence Morlet]

Also there is an update to the community contributed rwolf command; rwolf2. It has a few improvements & fixes