The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Child Health » How to combine data from different countries for a period of years?
How to combine data from different countries for a period of years? [message #3894] Mon, 02 March 2015 10:50 Go to next message
kinsukmanisinha@gmail.com
Messages: 9
Registered: January 2015
Location: Milan
Member
Dear all,

I am using the DHS database for the first time and also I have never worked with survey data before. My aim is a bit broad, it is to understand which variables have an impact on child mortality/under five mortality.

In order to achieve my aim I need to combine datasets from different countries for different year and build a pooled database.

And, this is the first problem I face -- I don't know how to combine databases from different countries for different year. Do I need to add a weight? If yes, how do I calculate this weight and how do I take it into account? Or take another factor into account.

I do understand that my question is a bit naive but if anyone could point me in the right direction, that would be simply great..!!!!

Many thanks..!!!

Regards
Kinsuk
Re: How to combine data from different countries for a period of years? [message #3916 is a reply to message #3894] Thu, 05 March 2015 16:00 Go to previous messageGo to next message
Trevor-DHS is currently offline  Trevor-DHS
Messages: 787
Registered: January 2013
Senior Member
1) If you are using DHS data for the first time and have never used survey data before, then you are probabbly taking on too much. I suggest learning to use a single survey dataset first and being able to reproduce published results before you try to do anything as complicated as you are suggesting.
2) It is not clear that you need to combine datasets based on the aim you have stated. You can analyze each dataset individually.
3) You should know that each individual dataset includes its own weight so you should be weighting data even if you are only analyzing one survey.
4) If you do need to pool data, you will need to denormalize the weights and then use the denormalized weights. There are several posts on the forum related to this.
5) Once you have the denormalized weights for your pooled data then you would apply weights just as you do for any other analysis, depending on the software you are using.
Re: How to combine data from different countries for a period of years? [message #3953 is a reply to message #3916] Tue, 10 March 2015 08:52 Go to previous messageGo to next message
kinsukmanisinha@gmail.com
Messages: 9
Registered: January 2015
Location: Milan
Member
Many Thanks for the detailed reply.

I agree with you, I should start with one data set first.

I read your points and I have a question:-

Point 3 says:-
3) You should know that each individual data set includes its own weight so you should be weighting data even if you are only analyzing one survey.

So, for example if I use only Egypt for year xxxx, I don't think I need to use weights. But if I use Egypt and Sudan for the same year, same survey, then I need to worry about weight. Is this what you meant?

Thanks once again..!!!
Re: How to combine data from different countries for a period of years? [message #4002 is a reply to message #3894] Mon, 16 March 2015 19:46 Go to previous messageGo to next message
Trevor-DHS is currently offline  Trevor-DHS
Messages: 787
Registered: January 2013
Senior Member
If you are using only Egpyt xxxx survey you should be using weights. The reason for this is that the probability of selection of households as used in the sample design varies for different parts of the country. To correct for this it is necessary to use weights. You will find the relative weights in one of the following variables:
v005 - for women and children's datasets
hv005 - for household and household member datasets
mv005 - for men's datasets

If you are combining surveys for, say, Sudan and Egypt, then you have to not only use the weights, but also denormalize the weights (described in other posts on the forum) before pooling the data.

[Updated on: Mon, 16 March 2015 19:50]

Report message to a moderator

Re: How to combine data from different countries for a period of years? [message #4007 is a reply to message #4002] Tue, 17 March 2015 06:50 Go to previous message
kinsukmanisinha@gmail.com
Messages: 9
Registered: January 2015
Location: Milan
Member
Thanks once again..!!!
Previous Topic: how to use svy with fisher's exact
Next Topic: From questionnaire to the variable
Goto Forum:
  


Current Time: Fri Mar 29 08:34:35 Coordinated Universal Time 2024