Dear all,
I am working with a dataset of 150 patients, where 179 outcomes (brain areas) are measured at two timepoints (pre-post).
A reviewer has requested an adjustment for p-values due to multiple testing.
I have never adjusted for such a large array of p-values, and I am not well-informed about the correlation between the tests.
I used a permutation based test (the Romano-Wolf procedure), but it seems to be overly conservative.
I read somewhere that a good alternative might be the method of Benjamini, Krieger, and Yekutieli.
What do you think? What would you do in my place?
Do you know if there is an STATA command that implements this method? I haven't found one yet.
Thanks in advance for your advice.
Gianfranco
I am working with a dataset of 150 patients, where 179 outcomes (brain areas) are measured at two timepoints (pre-post).
A reviewer has requested an adjustment for p-values due to multiple testing.
I have never adjusted for such a large array of p-values, and I am not well-informed about the correlation between the tests.
I used a permutation based test (the Romano-Wolf procedure), but it seems to be overly conservative.
I read somewhere that a good alternative might be the method of Benjamini, Krieger, and Yekutieli.
What do you think? What would you do in my place?
Do you know if there is an STATA command that implements this method? I haven't found one yet.
Thanks in advance for your advice.
Gianfranco
Comment