Lloyd Dumont <[email protected]>:
I would think it has more to do with how your data are -svyset-. If
you provide some info on the survey design, you might get a better
answer, but I would define strata distinctly across rounds, make sure
weights are scaled appropriately (do you want to account for change in
pop size?), then -svy: regress- on an indicator for round. Or -svy:
tab- versus round for indicator variables.
On Wed, May 21, 2008 at 7:24 PM, Lloyd Dumont <[email protected]> wrote:
> Hello, everyone. I wanted to examine a few variables
> for their differences-in-means over two rounds of a
> survey. That is, for each variable, how has its mean
> changed between survey round I and survey round II?
>
> I can imagine how I would undertake the t-tests if
> these were two, disjoint cross-sections or if these
> were a balanced panel. But, this is something in
> between. Some respondents are only in the first
> round, some are only in the second, and some were
> actually surveyed twice.
>
> Since my data are xtset, I would guess there is a
> straightforward way to do this.
>
> Thanks for your thoughts, of course. Lloyd Dumont
*
* 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/
