Rare cases of time varying psu and strata
Dear support team,
When trying to account for the complex survey design, I have run into the following stata error message: "panels are not nested within clusters". In trying to find the cause of the problem, I have noticed that there is one instance in indresp where psu/strata switch from 3924/2962 in wave 7 to 120/30 in wave 8 onwards, and one instance where psu/strata switch from 2918/2459 in wave 7 to 132/34 in wave 8 onwards. As I have read that with regards to "the strata and psu variables (...), as these are sample design variables, their values do not change across the waves for any person" (Introduction to Understanding Society using Stata, Example 6), this surprised me. Any clarification on why this is the case and how I should deal with it would be greatly appreciated.
PS: The stata code I used to examine the relevant psu/strata is as follows:
by pidp (psu), sort: gen byte moved = (psu1 != psu[_N])
browse if moved