Missing women's weight and height measurements in DHS-8 [message #29958] |
Thu, 29 August 2024 11:49 |
andreas85
Messages: 1 Registered: August 2024
|
Member |
|
|
I am working with the latest release of DHS-8. Across countries, I find that women's weight and height measurements are missing for a large share of the observations. For example, in the Kenyan individual recode file, v447 indicates "no measurement found in household" for 47.44% of the dataset. I cannot find an obvious reason for that, which is why I am concerned that the missing values might indicate a selection problem.
|
|
|
Re: Missing women's weight and height measurements in DHS-8 [message #29982 is a reply to message #29958] |
Thu, 05 September 2024 08:31 |
Janet-DHS
Messages: 878 Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member, Tom Pullum:
This survey included subsampling. If you go to the PR file and enter "lookfor select" you will see a variable hv027, "selected for male interview". This 0/1 variable refers to much more subsampling than just selection for the male interview. I suggest that in the PR file you enter "gen height_measured=1" and "replace height_measured=0 if ha3==.", and then "tab hv027 height_measured". This will show that in households with hv027=0, women's anthropometry was skipped. As I said, several other topics were also skipped in these households, not just the interviews with men.
In general, when large numbers of cases have a dot, for Not Applicable (NA), it is because of subsampling or some skip in the questionnaire. You can get more details by checking the skips in the questionnaire and looking at the first chapter and the appendix on sampling in the final report.
|
|
|