I conducted a national mail survey of physicians using stratified random
sampling and have a few Stata related analysis questions that I hope someone
on this list can help me with. Two groups of physicians were randomly
selected from a national physician database (internists and
gastroenterologists [GI]). Below is what we did with some made up numbers to
make it simple.
Internists: there are 10,000 internists in the US and we randomly sampled
5,000. A total of 2,500 of the 5,000 surveys were returned.
GI: there are 5,000 GI doctors in the US and we randomly sampled 1,000. A
total of 500 of the 1,000 surveys were returned.
I have data on non-responders for age, gender, and geographic location.
There were no significant differences between responders and non-responders
for these variables. My questions are as follows:
1) For internists, the base weight is 2 because we randomly sampled 5,000 of
the 10,000 physicians. Since we received 2,500 surveys back and there we no
differences between responders and non-responders, is it correct to use a
weight of 4 (2,500 internists returned completed surveys?
2) When I set up the survey analysis in Stata, I will have 1 variable for
strata (there are 2 strata - internists and GI) and a 2nd variable for
weight. Is that correct?
Thanks so much for your help.
Ed
*
* 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/
