Home » Data » Sampling » Two or Three Sampling Stages
| Two or Three Sampling Stages [message #1710] |
Tue, 01 April 2014 14:13  |
lukassg
Messages: 12
Registered: January 2014
|
Member |
|
|
Hey!
Could someone explain me exactly what is the difference between three-stage and two-stage sampling?
I very well understand the two-stage sampling:
1. I select a certain number of enumeration areas based on a previous census or national survey.
2. I sample a certain number of households from these EAs/clusters.
For the three-stage sampling:
As I understand it, this stage is before the EAs/clusters are selected. So we have even larger areas in the first stage, from which we then select the EAs/clusters in the second stage and the households in the third stage? Why do we have this additional stage in the beginning?
Finally, I am pooling always two different DHS survey for one country to get an over-time estimator of certain effects. I know that I have to adapt the stratification and PSU specifications. Is that enough in the case of a country, e.g. Tanzania, where I pool the years 1999 and 2010, where one year has 3 stages and the other 2 stages? Or does that make it impossible to analyse?
In this context, I also stumble upon probability proportional to size (pps). Can someone explain me this concept as well? :)
Thanks for your help!
Lukas
[Updated on: Tue, 01 April 2014 15:07] Report message to a moderator |
|
|
|
Current Time: Sat Nov 23 13:16:11 Coordinated Universal Time 2024
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
