I have a panel dataset of medical groups and their measurements of 10 time
occasions. The medical groups contains 135 intervention groups and 25
control groups.
Now I want to resample 200 replications (for modeling, predictions, and
then bootstrap the means).
1. In order to have enough sample size for the control group (for
modeling), I will need to stratify on this group variable, correct? And I
know that bsample has strata to achieve this.
2. However, as the plan is if a medical group is chosen, all of the time
occasions should be chosen. It should be clustered on medical group, right?
If i am right about the above two (strata, and cluster), I need to know
whether bsample allows both strata and cluster. Can someone help me on
this - whether Stata can do this or not, and if it can, what will be the
syntax?
