Hello, Statalisters;
Partly out of curiosity (and partly out of necessity), I would like to
ask a question regarding how people analyzed survey data with weight
(unequal selection probability) in the past. I tried several internet
search engines but never came up with any.
When I was a graduate student in the last century, I received a piece of
advice concerning how to use weight variable in survey data analysis.
The advice was "when the data set comes with an weight variable, you
should use it for descriptive purpose. But for regression analysis, you
don't have to use weight as long as your model is correctly specified."
Personally I do not subscribe to this idea (for practical purpose) but
can understand why some people said so. And I found a few articles in
my field (well, in political science) that ignored weight while
analyzing survey data.
Then, my question is whether ignoring weight (for regression) is a
common practice in some part of the discipline in the past or not. And
if so, I would like to know if there is any paper or reference that
explains why one can ignore weight for regression but cannot for
descriptive inference.
I apologize in case such advice or practice has never been endorsed.
It can be my false recollection of passing conversations.
Yours,
Yukio
*
* 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/
