The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » Other countries » Prevalence of stunting, wasting and underweight in BDHS 2014
Prevalence of stunting, wasting and underweight in BDHS 2014 Sat, 10 September 2016 12:38
 brammaputram@yahoo.fr Messages: 7Registered: November 2015 Location: Bangladesh Member
Hi,
I am working with BDHS 2014 BDPR datasets to measure the prevalence of stunting, wasting and underweight by using percentile scores of Height/Age, Weight/Height and Weight/Age. My findings are: 3%, 7%, 70% respectively, which are very dissimilar compared with those from BDHS 2014 final report, (which used z scores): 36.1%, 14.3%, 32.6%.

So far as I am concerned, percentiles lower than fifth should correspond to <2SD of the WHO Child Growth Standards median.

I tried calculating stunting by z-scores as well (HC3-HC6/HC5), but the result is still very different (13.4%).
Sorry if I am being dull here methodologically, but would really appreciate if you could provide some insight to it.

Re: Prevalence of stunting, wasting and underweight in BDHS 2014 [message #10791 is a reply to message #10772] Tue, 13 September 2016 12:23
 Bridgette-DHS Messages: 2307Registered: February 2013 Senior Member
Following is a response from DHS Senior Research Associate, Shireen Assaf:

To obtain the estimates of child stunting, underweight and wasting you need to use the PR file and select for de factor children. In addition, the z scores in the data file are multiplied by 100. The following code will give you the correct estimates.

```use BDPR70FL.dta, clear
gen wt= hv005/1000000

gen stunt=0 if hv103==1
replace stunt=. if hc70>=9996
replace stunt=1 if hc70<-200 & hv103==1
ta stunt [iw=wt]

gen under=0 if hv103==1
replace under=. if hc71>=9996
replace under=1 if hc71<-200 & hv103==1
ta under [iw=wt]

gen wast=0 if hv103==1
replace wast=. if hc72>=9996
replace wast=1 if hc72<-200 & hv103==1
ta wast [iw=wt]
```
Re: Prevalence of stunting, wasting and underweight in BDHS 2014 [message #11758 is a reply to message #10791] Tue, 07 February 2017 00:50
 brammaputram@yahoo.fr Messages: 7Registered: November 2015 Location: Bangladesh Member
Dear Shireen Assaf,

Thanks indeed for the clarification. Sorry I am following up very late.

I am using SPSS for my analysis so need to do the sorting manually without using the codes. From your text I have one more query, am I right in thinking that I can get the original z score just by diving by 100 since they were multiplied by 100 in the PR datasets. or there are any other procedures?

Re: Prevalence of stunting, wasting and underweight in BDHS 2014 [message #11773 is a reply to message #11758] Tue, 07 February 2017 12:07
 Bridgette-DHS Messages: 2307Registered: February 2013 Senior Member
Another response from Shireen Assaf:

Yes, to get the original z-scores you divide by 100. In the Stata code I provided for calculating stunting, I use hc70<-200 but this corresponds to a z-score<-2 .
 Previous Topic: Peru 2008-2012 and Juntos program Next Topic: Peru district/province variable 1992 svy
Goto Forum:

Current Time: Sat Jun 19 04:49:59 Coordinated Universal Time 2021