| Issues with weight and height data [message #16199] |
Wed, 21 November 2018 19:33  |
sdas@gradcenter.cuny.edu
Messages: 2
Registered: September 2018
Location: United States
|
Member |
|
|
Hi,
I have merged all the four rounds of the NFHS Individual Recode Data and just looking at the descriptive statistics through the years after some cleaning. The data for birth-weight, measured weight and height for children looks very noisy. I did checked the documents and removed the outliers from the data. Yet, the descriptive statistics looks very noisy. For instance, for the birth weight data of the first child, the data is clustered around round values like 2000,2500,3000, etc. and I cannot find any apparent reason why this is happening. Is the use of birth-weight, measured child weight and height credible for any analysis for all the 4 rounds of the NFHS? I know about a paper by Rohini Pande where she used the measured height of children for the analysis, but that was using only NFHS 3. It would be really helpful if anyone working with the same indicators can help me out with the issue.
Best,
Sagnik Das
PhD Student, CUNY: The Graduate Center
Sagnik
|
|
|
|
| Re: Issues with weight and height data [message #16230 is a reply to message #16199] |
Thu, 29 November 2018 09:36  |
fredarnold
Messages: 45
Registered: March 2014
|
Member |
|
|
|
There is always considerable heaping on birth weights, particularly for mothers who cannot show a written record of the birth weight. Many women do not know the exact birth weight, and rounding on values such as 2,000, 2,500, and 3,000 is common. This may be particularly true for mothers with a low level of education. Accurate reporting of the birth weight among mothers who were not able to show a written record of the birth weight is particularly difficult for births that occurred further in the past (up to five years before the interview).
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
