Merging IR to PR [message #26245] |
Thu, 23 February 2023 13:05 |
azimkl10
Messages: 3 Registered: January 2023
|
Member |
|
|
Hi,
I managed to merge the 2018 IR file to the PR file by merging with these variables ha0, hv001, and hv002. However, I want each child (5-15) to have their mother's information. The information I want from the IR are v106, v130, v481, v501, v505 merged into the children's data. For example, v481 is whether respondent (mother) is covered by health insurance. I want whatever the mother's answer is to that question and apply it to all her children if that makes sense. I also wonder how I can tackle this situation if there are multiple women in the household?
Additionally, there are some variables in the PR file which are mother specific like ha66 which is woman's highest education level which I want to apply to all her children (instead of leaving it as NA in the data set) and I want to know how to go about that?
The reason for doing this is because I plan to analyze children 5-15 (while using some mother specific variables as covariants) and I want some of the mother's information to be kept and assigned to their children. I will filter through ages 5-15 to conduct analysis and I do not want to lose their mother's information when filtering.
Thank you for the help! I am open to solutions in Stata and R.
|
|
|
Re: Merging IR to PR [message #26252 is a reply to message #26245] |
Mon, 27 February 2023 08:39 |
Bridgette-DHS
Messages: 3203 Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS staff member, Tom Pullum:
It's good that you want to study children age 5-15 (or 5-14, which would be more standard; at age 15, household members are eligible for the individual interviews). That age interval is under-studied.
If you want to link these children with their mothers, you are reducing to children who are in the same household as the mother, and if you want the mother's IR data, you are reducing further to mothers whose current age is 15-49. Some mothers of older children will themselves be too old for the individual interview. Those reductions may introduce some bias.
You could start with the BR file, which has one record for each child in the birth histories, and each record includes virtually all of the information on the mother in the IR file. Restrict to children whose age (b8) is in the desired range and who are living with the mother (b9=0) This file will only be missing the information about the child that is in the PR file. You would get that information by merging v001 v002 b16 in the BR file with hv001 hv002 hvidx in the PR file. I think this would be more direct than what you are describing.
Recently I posted a Stata program (see attached) that uses children age 0-17 as cases and merges all the information about them, their mothers, and their fathers, onto a single record. You could adapt that program for your purposes, cutting out the merge with the father's data.
[Updated on: Mon, 27 February 2023 08:39] Report message to a moderator
|
|
|