| Bangladesh 2011 DHS: Merging "children" and "male" files [message #1657] |
Mon, 24 March 2014 15:09  |
skatumuluwa
Messages: 1
Registered: March 2014
|
Member |
|
|
Hi,
I am working with the Bangladesh 2011 DHS dataset and am trying to merge the "children" and "male" files and have run into some difficulties.
Basically, I cannot figure out how to create a unique variable in both files so that I can merge based on that variable. Details follow.
Since the "male" file contains all the variables needed to re-create the women's CaseID (cluster + household + woman's line number) that is found in the "children" file, I went ahead and created a numeric variable in the "male" file combining these 3 variables. Since SPSS cannot merge files based on a string variable, I also converted the CaseID provided in the "children" file from a string to a numeric variable, which removed the spaces between the numbers in the CaseID. When I sorted by the woman's CaseID in both of my files and tried to merge the datasets, I got an error message saying that there were duplicate cases. When I went back to the data I realized that there are 2 ways to get each CaseID: example below
CaseID V001(cluster) V002 (household) V003 (respondent's line number)
1882 1 88 2
1882 18 8 2
I also tried merging the data sets based on the 3 individual variables (V001, V002, and V003) but you can only sort the data by one variable. When I sorted by 1 variable and then tried to merge the data based on the 3 variables, Stata gave me an error message saying that the variables had not been sorted.
Any suggestions of how to get around this dilemma?
Many thanks!
[Updated on: Mon, 24 March 2014 15:43] Report message to a moderator |
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
