The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Sampling » Two or Three Sampling Stages
Re: Two or Three Sampling Stages [message #2039 is a reply to message #2035] Wed, 16 April 2014 17:50 Go to previous messageGo to previous message
Trevor-DHS is currently offline  Trevor-DHS
Messages: 805
Registered: January 2013
Senior Member
To user-rhs:
I think the weights may be ok if you have two samples from the same country that have roughly the same sample size. If the sample size differed substantially between the two surveys or you are using data from two different countries then I would say that you need to adjust. Also note the caveats in Tom's note referenced below.

I rarely use a household N, but if you wish to you again have a number of possibilities. Some DHS samples are prepared with households as the measure of size and the total number of households may be reported in appendix A of the DHS report. If the measure of size was total population instead of households you could use the total population reported in the DHS report divided by the total de jure sample population from the dataset. You can also get an estimate of the total population from the UN web site that I referred to earlier. Counts of households aren't reported as widely as population counts, but there are some sources. For example, Wikipedia even has total households ( http://en.wikipedia.org/wiki/List_of_countries_by_number_of_ households). The source for this appears to be the UN Demographic Yearbook ( http://unstats.un.org/unsd/demographic/products/dyb/dybcensu sdata.htm (look for "households by type of household").

Also, I should mention an alternate source for population numbers - the US Census Bureau's International Data Base: http://www.census.gov/population/international/data/idb/info rmationGateway.php

To lukassg:
1) I don't think there is a right or wrong answer. I think you have to decide which is the most appropriate. You could decide to use the 2010 number as that is close to 2009, you could use the 2007 census number, or you could interpolate between the UN's 2005 and 2010 numbers. I don't think it will really matter in an meaningful way. Just document what you actually do in your analysis writeup.

2) You don't need to re-normalize weights - it will have no effect on the results of your analysis, other than producing much bigger Ns.

I hope this helps.

Report message to a moderator

[Message index]
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: Stratification
Next Topic: Enumeration Area size in BDHS
Goto Forum:
  

[ Syndicate this forum (XML) ] [ RSS ] [ PDF ]

Current Time: Mon May 19 00:57:53 Coordinated Universal Time 2025
.:: Contact :: Home :: Privacy Policy ::.

Powered by: FUDforum 3.0.9.
Copyright ©2001-2018 FUDforum Bulletin Board Software

The information provided on this Web site is not official U.S. Government information and
does not represent the views or positions of the U.S. Agency for International Development or the U.S. Government...