Hi - Say you have a list of unique IDs.
That list is forwarded to some service vendor that processes a subset of that list, with the assurance given to you that the subset in question will be randomly drawn from your list.
How can I test that? A simple n-th selection is easily ruled out, but how to feel comfortable the selection was properly randomized?
If I have other variables associated with the IDs, I can run t-tests with the flagged processed IDs, but this time around it's only that single ID variable.
Say they process 70% of the list: my approach is to bootstrap the proportion and test p = 0.70, keeping an eye on any activity away from the mode/mean.
Is there a NPAR TESTS option or other nonparametric approach to look at this problem?
Thanks.
Marc.
It’s the same Hotmail®. If by “same” you mean up to 70% faster.
Get your account now.
Locked
