Although this is technically correct, analysts often take the simplifying
approach of using one weight for analysis that is the product of design
(selection-probability) weights, selection-specific nonresponse weights, and
sample-to-population calibration (post-stratification) weights. Some
software allows partitioning of these weight factors. The difference in
estimation is usually negligible.
Arnold H. Levinson, Ph.D.
Associate Scientist
AMC Cancer Research Center
Assistant Professor
University of Colorado School of Medicine
303.777-8801
[email protected]
Date: Thu, 18 Sep 2003 10:39:01 +0100
From: "Patrick Sturgis" <[email protected]>
Subject: st: nonresponse weights
Someone has told me that the svy commands in stata can provide correct
variance estimates for design weights but non for nonresponse (callibration)
weights. Does anyone know more about this? Thanks in advance,
Patrick
*
* 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/
