BMI in PR vs IR File [message #29126] |
Fri, 26 April 2024 09:34 ![Go to next message Go to next message](/theme/default/images/down.png) |
AkhilK28
Messages: 19 Registered: February 2022
|
Member |
|
|
Hello!
When calculating BMI in the PR vs. IR file for NFHS-5, the PR file has more observations for women aged 15-49 years old.
In the PR file, I am using ha variables to correspond to the biomarker questionnaire -- see the code below:
keep if inrange(ha1,15,49)
keep if inrange(ha40,1200,6000)
Observations with valid BMI Values for 15-49 year old women in PR file: 709,252
In IR file, I am using v445:
keep if inrange(v445,1200,6000)
Observations with valid BMI Values for 15-49 year old women in IR file: 699,362
Can someone at DHS let me know why this difference exists?
What is official guidance? Stay to biomarker in PR? But why a difference of ~10,000?
[Updated on: Fri, 26 April 2024 09:35] Report message to a moderator
|
|
|
|
|
Re: BMI in PR vs IR File [message #29228 is a reply to message #29218] |
Mon, 13 May 2024 11:28 ![Go to previous message Go to previous message](/theme/default/images/up.png) ![Go to next message Go to next message](/theme/default/images/down.png) |
Bridgette-DHS
Messages: 3151 Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Staff Member, Tom Pullum:
All of the women in the IR file are also in the PR file, and all of the men in the MR file, are also in the PR file. Height and weight are collected as part of the household survey, not as part of the individual surveys of women and men, so there are a few (usually just a few, sometimes many) additional cases in the PR file that are not in the IR or MR files. I think the decision about whether to use the PR file for the BMI scores depends on what covariates you want to include in the analysis. If you will only use household-level covariates, then use the PR file. But if you want to include variables that come only from the individual surveys, use the IR (or MR) files.
The same applies for children and Z scores (the HAZ.WZS, and WHZ). You can get them from the PR file or the KR file, with a few more cases in the PR file than in the KR file. If you want to use covariates that are specific to the child or mother, use the KR file. For covariates that are household-specific, use the PR file.
But I'd advise against switching from one file to another within the same report, because that will confuse the user. Often it will be better just to use the IR or MR or KR file.
|
|
|
Re: BMI in PR vs IR File [message #29235 is a reply to message #29228] |
Mon, 13 May 2024 17:51 ![Go to previous message Go to previous message](/theme/default/images/up.png) |
AkhilK28
Messages: 19 Registered: February 2022
|
Member |
|
|
Thank you so much, Tom, this is very helpful!
We are planning to go with the IR file since that has some self-reported covariates from the women questionnaire.
Thank you once again for your help :)
|
|
|