[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: Methodological issue: combining datasets from different populations

From   "Martin Weiss" <>
To   <>
Subject   st: RE: Methodological issue: combining datasets from different populations
Date   Mon, 6 Jul 2009 21:18:33 +0200


This does not strike me as a homework question / I am willing to err for the
benefit of Georgiana. She might want to look at -suest- which is an
underappreciated command and happily works with -svy- commands. The help
file gives examples of this kind of usage...


-----Original Message-----
[] On Behalf Of Georgiana Bostean
Sent: Montag, 6. Juli 2009 19:24
Subject: st: Methodological issue: combining datasets from different

Hello everyone. I have a statistical/methodological question for you. Any
suggestions or thoughts are welcome.

I am combining (appending) two separate nationally-representative datasets
(drawn from two different populations, two countries). I am interested in
comparing certain outcomes of groups within those populations.

A hypothetical example: A nationally- representative dataset from Germany
and a nationally-representative datatset from Turkey. I want to compare
Turkish residents in Germany to Turks in Turkey along certain (dichotomous
and ordinal) outcomes, controlling for a number of factors. Thus, the
independent variable (migration status) would have categories that are
exclusively drawn from one or the other population. The datasets have
comparable questions for the outcome, so that is not the issue.

But, I have a hunch that merely doing a logistic or ordinal regression
would be wrong.

Any suggestions about which direction to take this? First, about what type
of analysis would be appropriat and second, how to handle each dataset's
weights (pweight, psu, strata for each of the two datasets)?

Thank you in advance for any suggestions!

Georgiana Bostean
Department of Sociology
University of California, Irvine

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index