|
|
FPC with Systematic PPS Sampling |
|
Message/Author |
|
|
Do the standard instructions for the TYPE=COMPLEX, FINITE option still hold when the primary sampling units are sampled using systematic probability proportional to size sampling? For example, if I want to use the POPULATION setting, then I would normally specify: FINITE is popsize (POPULATION); where popsize is a variable containing the number of clusters in the population for a given stratum. So I'm wondering, does "popsize" have the same meaning under systematic PPS sampling? As a related question, is the FPC even appropriate with systematic PPS sampling? Theoretically, a PSU can be selected more than once under this algorithm. Thank you! |
|
|
I don't see any issues with PPS, it is just like any other unequal probability sampling and can be done WOR sampling. This article gives details on the Mplus implementation https://www.statmodel.com/download/SurveyJSM1.pdf |
|
Back to top |
|
|