version 13.1 * svy command -- note that many commands deliberately produce * fatal errors, Also ftest must be installed from SSC * The svy and svyset commands webuse nhanes2f, clear svyset sum finalwgt stratid psuid reg weight height age i.female i.black svy: reg weight height age i.female i.black * Subsample analyses svy, subpop(if female==0): reg weight height age i.black * Commands that don't work & their alternatives webuse nhanes2f, clear svy: summarize diabetes black weight height svy: mean diabetes black weight height sw, pe(.05): svy: logit diabetes black weight height * svy post-estimation commands webuse nhanes2f, clear svy: mean female black height weight diabetes estat sd * OLS tests that do and do not work svy: reg weight height age est store m1 svy: reg weight height age i.female i.black est store m2 test 1.female 1.black ftest m1 m2 nestreg, quietly: svy: reg weight (height age) (female black) * OLS post-estimation that does not work quietly svy: reg weight height age female black dfbeta estat hetttest estat imtest rvfplot estat lvr2plot * CDA models -- What works and what does not webuse nhanes2f, clear * Constrained model svy: logit diabetes female black est store m1 svy: logit diabetes female black weight height est store m2 lrtest m1 m2, all * Use Wald test instead test weight height