Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: permutation, strata() for empirical statistic distributions


From   David Airey <david.airey@vanderbilt.edu>
To   statalist@hsphsun2.harvard.edu
Subject   st: permutation, strata() for empirical statistic distributions
Date   Wed, 5 Nov 2003 07:51:46 -0600

If you have a model, and you wish to determine the multiple test probability for your data set, you can use a multiple test correction procedure, or you can permute the data set many times to get an empirical distribution of test statistics.

(1) I'm assuming if the data is grouped or stratified, the permutation procedure should respect that. So if student|class|school, then you don't want to permute the student into a class he was never in. I see an option in permute that is "strata". Does this serve this purpose?

(2) A permutation procedure that respects nesting, should have a model that respects nesting. That is, it makes no sense to respect the nesting in the permutation procedure, but then ignore strata in the model, right?

Thanks,

Dave

*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/




© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index