The DHS Program User Forum      
Discussions regarding The DHS Program data and results
Home » Data » Merging data files » Linking GPS data to DHS data files
Linking GPS data to DHS data files [message #12910] Sat, 05 August 2017 13:33 Go to next message
davidk35 is currently offline  davidk35
Messages: 1
Registered: August 2017
Hi everyone,

I am currently creating a panel dataset for African countries, which uses years and administrative regions at the first level (as given by gadmn) as the units of analysis.
As I would like to compute for example the share of young men (15-24) to the adult population (e.g. until 54) in a specific region and year, I need to assign each cluster mapped by the GPS file to the corresponing cluster in the DHS file (e.g. GNGE52FL and GNMR52DT).
I thought I could do this by the DHSID but it turns out that the DHSID is not equivalent to any combination of variables in the DHS file (e.g. I thought of creating a string in STATA with the code egen = concat (mv000 mv007 mv001) to create the DHSID, but the result is not eqivalent to the DHSID in the GPS-file wich is GN20050000001 for the first cluster for example).

therefore my question would be how to match/merge the DHS-file with the GPS-file such that I can do regional analysis using the clusters in this region for a given (survey) year?

Thank you very much for your help in advance.
Re: Linking GPS data to DHS data files [message #12916 is a reply to message #12910] Mon, 07 August 2017 16:37 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 1112
Registered: February 2013
Senior Member

Following is a response from Senior DHS Specialist, Trinadh Dontamsetti:

The DHSID column you are referring to does not have a link to any of the variables you mentioned. The DHSID GN20050000001 is simply the country code (GN), survey year (2005), and the cluster number (001), spaced with placeholder zeroes. To link the GPS data file to the survey dataset, you will need to reference the DHSCLUST column in the GPS file, which should match v001 or hv001 from the recode. Doing so should allow you to match the two datasets.
Previous Topic: Combining PR and KR to get child-specific malaria test results
Next Topic: Merging the child, household and wealth file for Philippines 1998
Goto Forum:

Current Time: Tue Oct 17 05:46:07 Eastern Daylight Time 2017