| Implausible waist and hip circumference values in NFHS-5 data [message #26947] |
Wed, 31 May 2023 05:50  |
MRedwanul
Messages: 8
Registered: December 2020
|
Member |
|
|
Hi,
We have been exploring the data on waist and hip circumferences (among 15 to 19 years old), which appear to have been measured in centimeter. However, when we look at histograms, we find values that are too large to be plausible.
In both of the variables, there are a number of "999.5" values. Should we assume these values to be missing?
Could an expert also please tell us whether there are arbitrary values for waist and hip circumferences, that we can use to flag observations above those as implausible? For example, in waist circumference variable, there are values of 176 (cm) which correspond to , and is highly likely to be implausible
Many thanks in advance!
|
|
|
|
| Re: Implausible waist and hip circumference values in NFHS-5 data [message #26949 is a reply to message #26947] |
Wed, 31 May 2023 09:45  |
fred.arnold@icf.com
Messages: 84
Registered: May 2021
|
Senior Member |
|
|
The health investigators in NFHS-5 who conducted the hip and waist measurements were extensively trained in the correct measurement procedures. They also practiced the procedures during the survey pretest. All hip and waist circumferences were measured in centimeters. For those who were measured, the measurements were recorded in four boxes (three boxes plus a fourth box after the decimal point). For those who were not measured, the following responses were recorded:
REFUSED.............9995
OTHER...............9996
Therefore, you can assume that both 999.5 and 999.6 were missing cases.
Also note that hip and waist measurements included women who were pregnant, so those women should not be included in any tables for your analysis.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
