The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Weighting data » Svysey vs iweght
Svysey vs iweght [message #2863] Wed, 03 September 2014 10:33 Go to next message
jtdkelly@gmail.com is currently offline  jtdkelly@gmail.com
Messages: 11
Registered: September 2014
Member
Hello,

In the DHS user fora I've read, I have seen people use both the svyset and the weight commands in stata to analyze DHS data. For instance, I have seen both of the following:

tab var1 var2 [iweight=dweight], row
svy: tab var1 var2

Could you comment on the difference between these commands and when it is appropriate to use each?

Many thanks!

Re: Svysey vs iweght [message #2872 is a reply to message #2863] Thu, 04 September 2014 12:36 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3016
Registered: February 2013
Senior Member
Following is a response from Senior DHS Stata Specialist, Tom Pullum:

Three points that may help:
(1) In estimation commands, you can adjust for weights and clusters within the command, for example with "regress y x [pweight=v005], cluster(v001)", and you don't really need svyset and svy, although many people use them anyway. However, the only way to make the adjustment for strata is by using svyset and svy.
(2) Unfortunately, some commands in Stata such as tab do not allow pweights, in particular, so many people will use svyset and svy: tab (say) for such commands.
(3) Adjustments for clusters and strata will only affect standard errors. These adjustments should be used if you are estimating standard errors (for tests or confidence intervals) but are not relevant otherwise. The adjustment for weights will affect estimates AND standard errors.
Previous Topic: Variable Verification in Zim DHS 2010-11
Next Topic: Confusion re: "Numbers in parentheses are based on 25-49 unweighted cases"
Goto Forum:
  


Current Time: Thu Mar 28 10:08:24 Coordinated Universal Time 2024