Re: Re: st: Combining a survey weight and a frequency weight

["Ariel Linden, DrPH" <ariel.linden@gmail.com>]

I am not sure where you "found out earlier that the weight used in matching
is to tell how Stata how important the matched observation is...", but that
is not the case (nor do I understand what that really means).

In k:1 matching, weights are given to controls in order to ensure that they
sum up to the value of the matched treated unit. So if you have 4 controls
for each treated subject (which receives a weight of 1 in an ATT estimator),
each control receives a weight of 0.25, so that the sum = 1. So this is not
consistent with your statement: "how important a matched observation is".

I am not sure exactly what you would call this weight when it is multiplied
by a survey weight, but for units that receive a weight of 1 (all treated
units, and those matched controls where there is only one control matched to
the treated unit), the weight would reduce to the non-response weight that
you indicated earlier. For controls that have a different "matching weight",
this weight would account for both the number of controls matched to the
given treated unit as well as non-response...

I hope this helps

Ariel

Date: Sat, 24 Aug 2013 14:48:49 +0100 (BST)
From: James Jensen <jamesjensen212@yahoo.com>
Subject: Re: st: Combining a survey weight and a frequency weight

I would like to add that the frequency weight is generated by propensity
score matching by the psmatch2 command. I found out earlier that this weight
is a frequency weight that is to tell how important Stata how important the
matched observation is. 

I would like to combine this frequency weight and the survey weight when I
carry out the analysis on the matched sub-sample.

But I am not sure what the weight type should be.




*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/