Robert,
You can also try to run:
logit y x1 x2 x3 [pw=expansion_factor]
and see how many observations are being included, then
logit y x1 x2 x3, cluster(PSU)
logit y x1 x2 x3 [pw=expansion_factor], cluster(PSU)
where PSU are your primary sample units. Again you will check the number of
observations. This will help you identify if the missing values are coming
from expansion factors, PSUs or strata.
Rafa
----- Original Message -----
From: "Owens, Douglas Hamilton" <[email protected]>
To: <[email protected]>
Sent: Thursday, June 10, 2004 12:55 AM
Subject: st: svy commands and missing cases
> how about using e(sample) to identify the obeservations dropped from the
> svylogit? Is there something odd about the psu, strata, or weight in
those
> observations? Perhaps -svydes- might report something useful.
>
>
> > Hey Stata listers,
> >
> > I hope you can help me. I am stumped on what seems to be an easy
problem to
> > fix.
> >
> > I am trying to estimate a logit model using the survey commands on a
sample
> > of 9,966. When I estimate the model:
> >
> > svylogit y x1 x2 x3
> >
> > Stata lists 9,559 observations
> >
> > When I estimate the model without the survey commands:
> >
> > logit y x1 x2 x3
> >
> > Stata lists 9,966 observations, which is the correct number of
observations.
> > How can I get the svy procedures to estimate the model on ALL
> > observations?
> >
> > Thoughts, anyone?
>
>
> *
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
