Home » Data » Weighting data » Clustered Standard Errors
Re: Clustered Standard Errors [message #9278 is a reply to message #9219] |
Wed, 02 March 2016 11:57 ![Go to previous message Go to previous message](/theme/default/images/up.png) ![Go to next message Go to previous message](/theme/default/images/down.png) |
Bridgette-DHS
Messages: 3151 Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Stata Specialist, Tom Pullum:
Yes, if you construct svyset as you say, and then put "svy: " in front of an estimation command, you will get robust standard errors.
You can make the cluster adjustment in two ways. One is with svyset, the other is with regress y x, cluster(v021). (I think you meant to put "cluster(v021)" rather than "cluster(ADM1_CODE)".)
You can make the weight adjustment in two ways. One is with svyset, the other is with regress y x [pweight=v005].
You can also do regress y x [pweight=v005], cluster(v021).
You cannot, however, include the stratum adjustment within an estimation command. That adjustment can ONLY be made with svyset.
Note that v001 and v021 are exactly the same.
Something has gone wrong with v024 in your DRC 2007 data file. I expect that it has been recoded incorrectly. A conspicuous warning that something was wrong is the huge number of cases in the lines with labels.
I don't know which of the DRC 2007 files you are using, but here are the codes for the PR file
. label list hv024
hv024:
10 kinshasa
20 bas-congo
30 bandundu
40 equateur
50 orientale
61 nord-kivu
62 maniema
63 sud-kivu
70 katanga
80 kasa-oriental
90 kasa-occident
In the PR file, the total household population is 48,291. You are way off.
|
|
|
Goto Forum:
Current Time: Fri Aug 30 23:26:39 Coordinated Universal Time 2024
|