| Partners variables [message #8919] |
Thu, 14 January 2016 11:09  |
dhsid
Messages: 1
Registered: January 2016
|
Member |
|
|
Hi,
I have a question related to the Timor Leste DHS dataset.
I have merged the men's file to the women's file for currently married individuals.
To merge the two files I have used as key variables the following variables : v001 v002 mv034 (which corresponds to v003 in the women dataset)
I have checked that the partner's age and years of education variables included in the women dataset corresponds to the partner's age and years of education merged from the men's dataset . Apparently they do not seem to match. So I am wondering whether the merge is incorrect or whether the partner's related variables in the women dataset are not accurate.
Any help on this would be extremely useful.
thank you!
|
|
|
|
| Re: Partners variables [message #8961 is a reply to message #8919] |
Tue, 19 January 2016 15:16  |
Bridgette-DHS
Messages: 3214
Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Stata Specialist, Tom Pullum:
You have probably done the merge correctly. However, there is already a couple's file (a CR file) for all surveys that included a survey of men as well as women. You can check whether your file matches that one.
There are often discrepancies between a self-report of age, etc., and a partner's report of the same thing. Similarly, even in studies in which the same person is re-interviewed, they may give different reports of their own age. These are contexts in which some characteristics that are important in the West (age being the most notable example) are not really very salient in the everyday lives of people in traditional settings. Discrepancies do NOT mean that there is an incorrect match. (Occasionally, perhaps, but not typically.) You are ok.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
