| Re: issue about data merge : senegal 2012-2019 [message #27148 is a reply to message #27015] |
Tue, 20 June 2023 16:28  |
Janet-DHS
Messages: 893
Registered: April 2022
|
Senior Member |
|
|
Following is another response from DHS Lead Statistician, Ren Ruilin:
If I summarize the user's question, it is basically three questions:
1. Are the survey results comparable across survey years? The answer is yes, users can compare the indicators across surveys, normally at national level. Comparisions at sub-national levels are not guaranteed since the survey were not designed to do so.
2. Can the different survey years data be combined together? The answer is yes, but with conditions. One cannot combine directly the different years survey data because of the normalized weight. Before data combining, the sampling weight needs to be de-normalized. We had answered countless times the question on combining DHS survey data from the data user forum, the user can search on the data user forum to get the detailed instructions on how to combine survey data using normalized weight.
3. Are the different year's survey continuous survey or not? The answer is "it depends". I believe we have a small document accompanied to each data file explaining the data collection methodology. But this is not quite relevant with the user's aims of study. DHS always releases combined data if the different years data are from a continuous design. So, the user can judge based on the released data.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
