Dear statalisters,
I have data from two periods of a quasi-panel study. In the first period,
a clustered sample of certain communities and households within those
communities was taken. In the second period, a new sample was taken at the
same locations. So I have same clusters, but possibly (and most likely)
different households. I want to estimate the difference of a
characteristic, i.e., the change from the first period to the second, and
put an error on it. If I simply ignore the fact that those samples came
from the same location, I could use a version of a t-test with variances
corrected for clustering. But I expect the correlation over time to be
pretty high, as the characteristic is pretty stable, so the estimate of
the variance as the sum of variances, ignoring the correlations, would be
way too conservative. The question is:
(a) can I trick -_robust- or any of the -svy- commands into doing this for
me?
(b) if not, at the theoretical level, is there any reference I can look up
on this rather than deriving the estimate myself from the Taylor series
expansion?
--- Stas Kolenikov
-- Ph.D. student in Statistics at UNC-Chapel Hill
- http://www.komkon.org/~tacik/ -- [email protected]
* This e-mail and all attachments to it are not intended to provide any
* reasonable point of view and was transmitted to you in error. It
* should be immediately deleted by all recepients unless they really
* enjoy communicating with the author :). Other restrictions apply.
*
* 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/
