Ralph Mayrell wrote
> Also, we previously used the jointprob function in SUDAAN to implement a joint inclusion matrix; is there an equivalent in Stata? Or will the svyset commands accomplish the same ends.
Yes
Stata will not use the joint-inclusion probabilities, but it will
produce correct standard errors if you correctly specify the fpcs for
estimates of means, totals, proportions, and ratio (i.e. for
descriptive statistics) with linearization and jackknife variance
estimates ( but not for BRR, apparently). If you want to test for
associations of study factors, say with regression, you should
-svyset- without fpcs.
> Is it the psu identifier that Stata uses to identify a "certainly selected" psu; that is, if the psu identifier is the same for all observations in a strata, then Stata will treat it as a certainty PSU?
No. Stata recognizes that there is one sampling unit but it does not
know if there could have been others. It uses the fpc at the
appropriate sampling stage to decide whether the single unit is a
certainty unit (fpc =1) or not.
