I shall be grateful for comments by Statalist members survey design and model specification with a data set which was collected using stratified random sampling of persons in each of 50 U.S. states, with two strata ("stratum") per state, and includes a final respondent weight variable ("finalweight") for individuals sampled with the strata of each state. Specifically, thoughts are appreciated regarding:
a) the appropriateness of survey design setting as:
.svyset [pweight=finalweight], strata(stratum)
and
b) the stata model syntax in order to obtain: effect of state overall, the individual state-level coefficients, and coefficients for state-level measures; while controlling for individual-levels effects (e.g. age, sex, etc...)
0 Response to survey design and model specification
Post a Comment