Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: svyset

From   "Sayer, Bryan" <>
To   "''" <>
Subject   st: RE: svyset
Date   Fri, 21 Feb 2003 10:32:59 -0500

I don't quite understand how there are only two strata if the villages are
stratified by both income and geography.  It seems to me that there has to
be at least four, unless all of the villages in one geographic area are
poor, and all the ones in the other area are not-poor (or something like

That said, I think I would define the strata down to the village level and
specify the village as the PSU.  Although a case can be made for households
as PSUs, maybe.

Since you only have a household weight, then I guess you will need to use
that.  Unless you have enough information to create a person weight.

This is all based on the information here.  If anything is left out, the
suggestions could be incorrect.

Bryan Sayer
Statistician, SSS Inc.

-----Original Message-----
From: Zhonghe Li []
Sent: Friday, February 21, 2003 12:44 AM
Subject: st: svyset

Hi all,

In our survey, we have selected 2 townships from 1 county.  And then all the
villages in each township were stratified by average per capita income level
and geographic location into two strata.  Within each stratum of villages,
n1 villages were randomly selected.  Among the n1 villages selected, n2
households within each selected village were randomly selected;  and then
the household head was interviewed for household level information; and each
individual within the household was interviewed for their health status,
expenditure and so on.

My question is exactly how to set up the strata, cluster and the pweight in
this example.  For example, the pweight is really for the household level.
But when I am estimating the individual level variables, how should I
caculate the weight?  Should it carry the household level weight?

Thank you.

Elena Li

*   For searches and help try:
*   For searches and help try:

© Copyright 1996–2019 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index