Re: Child discipline merging [message #29611 is a reply to message #29602] |
Wed, 10 July 2024 09:48 ![Go to previous message Go to previous message](/theme/default/images/up.png) |
Janet-DHS
Messages: 816 Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member, Tom Pullum:
I have just looked at the HR and PR files for the Burkina Faso survey you mentioned. In this survey, at least, just one child was selected for this module and the variables for that child were coded in the HR file with prefix "hcdi" and no subscripts. In particular, hcdi2 is the household line number of the selected child.
In the PR file, which is a re-shaping of the HR file such that cases are individuals rather than entire households, the hcdi variables were copied onto every person in the household. This is potentially misleading for researchers who may think that the variables only appear on the record or line for the specified child, but for other purposes it can be helpful.
The line number in the household is hvidx. If you wanted a file of selected children, you would open the PR file and then enter "keep if hvidx==hcdi2". If you wanted to attach the variables to the mother, you would open the PR file and then enter "keep if hvidx==hv112". You would then merge that file with IR file, matching hv001 hv002 hvidx with v001 v002 v003 in the IR file. I think you have already done this or something equivalent.
Perhaps you can do what you want using hv101, the relation to head code. If you wanted to attach the hcdi variables to the household head, for example, you would open the PR file and enter "keep if hv101==1". But I am not sure how you intend to specify "the main female respondent". The household respondent is the person with hv003. If you open the PR file and enter "keep if hvidx==hv003" you will get a file of household respondents, with the ecdi variables on each record. The household respondent may or may not be a woman. You will have to settle on a definition, perhaps using a combination of hv101, hv104 (sex of household member), etc.
Finally, I would not be surprised if the coding of these variables is different in different surveys and may have evolved over time. I have worked with the early child development module but not with the child discipline module. The ECD variables appear with different names and other changes over time.
|
|
|