The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » India » ICDS pregnancy benefits coverage_ Mismatch with State reports
ICDS pregnancy benefits coverage_ Mismatch with State reports [message #18193] Wed, 09 October 2019 08:10 Go to next message
marian is currently offline  marian
Messages: 11
Registered: August 2019
Member
Hi,

I want to estimate the percentage of mothers who got any benefits under ICDS during their pregnancy (s562) but there are differences that are occurring in sample size with state nfhs 4 reports.

I have used filters such as children below 6yrs and children who are alive. However while for some states like Assam, I am getting exact estimates with state NFHS 4 reports, for others states (Kerela, Bihar) there are differences in both sample size and estimate.. Note, the sample size differences are between 5-20 observations of mothers of 0-71m children which results in 1-3 percentage point differences in estimates. I want to know if I have to generate a mother variable (or do a merge with v003) in BR file to get exact estimates across any state in India. There is no way to generate mothers who had a pregnancy ever. Please help.

[Updated on: Thu, 10 October 2019 03:29]

Report message to a moderator

Re: ICDS pregnancy benefits coverage_ Mismatch with State reports [message #18258 is a reply to message #18193] Mon, 21 October 2019 13:44 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 1748
Registered: February 2013
Senior Member
Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

You should use the BR file to access this variable all the way out to age 71 months. In this file, age in months is given by hw1, which is constructed as hw1=v008-b3 but only goes out to 59 months. I propose the following lines, after you have opened the BR file. You do not have to do any merging. Note that all of the KR variables are also in the BR file, but are coded NA (with a dot) for hw1>59. If you use the BR file but drop all cases with hw1>71, you can treat it as equivalent to a KR file with the addition of s555 for months 60-71; for that variable you would use the variable I call "hw1_to_71_months", which is hw1 extended to 71 months. (You could re-calculate hw1 itself to go out to 71 months but it is risky to recode one of the original variables and keep the original name.)


gen hw1_to_71_months=v008-b3
replace hw1_to_71_months=. if v008-b3>71

* confirm that this matches the original hw1 for months 0-60; a correlation omits cases that are NA on either variable 
summarize hw1 hw1_to_71_months
correlate hw1 hw1_to_71_months

tab hw1_to_71_months, summarize(s555)
Re: ICDS pregnancy benefits coverage_ Mismatch with State reports [message #18261 is a reply to message #18258] Tue, 22 October 2019 03:26 Go to previous message
marian is currently offline  marian
Messages: 11
Registered: August 2019
Member
Dear Mam,

Thank you.. It worked for me..
Previous Topic: Median number of years of schooling completed
Next Topic: Nutritional status of women
Goto Forum:
  


Current Time: Mon Feb 24 15:59:02 Eastern Standard Time 2020