The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Domestic Violence » Cluster identifiers and finite population correction
Re: Cluster identifiers and finite population correction [message #8692 is a reply to message #8644] Wed, 02 December 2015 14:48 Go to previous messageGo to previous message
Liz-DHS
Messages: 1516
Registered: February 2013
Senior Member
Dear User, Here is a response from Sampling Expert, Dr. Mahmoud Elkasabi:
Quote:
Most of the DHS surveys are based on a 2-stage design, in which sampling clusters are selected as PSU's in the first stage and households are selected in the second stage. To properly calculate the sampling variance, only the sampling clusters should be declared in the survey package. Therefore, I would say design (b) is the correct one
mydesign <- 
svydesign( 
id = ~V021, 
strata = ~V022 ,
data = data , 
weight = ~samplewt,
)

The sampling clusters can be id = ~V021 or id = ~V001. In most of the countries, both should be the same.

Regarding the fpc, for many reasons we do not provide the fpc value in the results report. Since we are dealing with national population surveys, the fpc value should be very small. Therefore, you can ignore the fpc in calculating the sampling errors and it should not affect the results.

 
Read Message
Read Message
Read Message
Previous Topic: Pakistan 2014 IPV analysis -- multi-level weights
Next Topic: Sexual/physical violence experience among men
Goto Forum:
  


Current Time: Mon Dec 6 13:16:49 Coordinated Universal Time 2021