The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Nutrition and Anthropometry » Slight difference in the Senegal Stunting, Wasting, Underweight prevalence (Slight difference in the Senegal Stunting, Wasting, Underweight prevalence result)
Slight difference in the Senegal Stunting, Wasting, Underweight prevalence [message #24371] Tue, 03 May 2022 09:27 Go to next message
tundexZ is currently offline  tundexZ
Messages: 4
Registered: March 2022
Member
Hi All,

I have tried to replicate the Senegal DHS Stunting, Wasting and Underweight results since (1986-2019) but it seem I keep getting slight different figures for some years. Is this results fine to work with? because I have gone through again all the merging process of all the datafile of the years together and yet I am still getting the same results or is there something I might have ommited?


Stunting (** indicates the results I got)
2019 DHS 17.9(17.7)** 2018 DHS 18.8(18.7)**, 2017 DHS 16.5 (16.5), 2016 DHS 17 (16.5)**, 2015 DHS 20.5 (20.4)**, 2014 DHS 18.7(18.7) 2012-13 DHS 18.7(18.4)**, 2010-11 DHS 26.5 (26.6)**, 2005 DHS 19.6 (19.3)**, 1992-93 (29.49) 1986 (28.21)

Wasting
2019 DHS 8.1 (8.1), 2018 DHS 7.8 (7.9)**, 2017 DHS 9(9.0), 2016 DHS 7.2(7.3)**, 2015 DHS 7.8(7.9)**, 2014 DHS 5.9(5.9), 2012-13 DHS 8.8 (8.7)** 2010-11 DHS 10.1(10.0)** 2005 DHS 8.5 (8.5)** 1992-93 (9.34), 1986 (5.13)

Underweight
2019 DHS 14.4 (14.1), 2018 DHS 15.1(15.1), 2017 DHS 14.4 (14.6)**, 2016 DHS 13.5 (13.5), 2015 DHS 15.5 (15.6)**, 2014 DHS 12.6(12.5)**, 2012-13 DHS 15.7 (15.7), 2010-11 DHS 17.7 (17.5)**, 2005 DHS 14.2 (13.7)**, 1992-93 (18.03), 1986 (16.99)


This are my codes for the analyis after merging all the dataset together

keep if hv103 ==1 & hc1 <=59
gen swt=hv005/1000000
gen haz=hc70/100
gen waz=hc71/100
gen whz=hc72/100
replace haz=. if haz>6
replace waz=. if waz>6
replace whz=. if whz>6
gen stunting=1 if haz<-2
replace stunting=0 if haz>=-2
replace stunting=. if haz==.
label define stunting 0"non stunted" 1"stunted",replace
label value stunting stunting
gen wasting=1 if whz<-2
replace wasting=0 if whz>=-2
replace wasting=. if whz==.
label define wasting 0"non wasted" 1"wasted",replace
label value wasting wasting
gen underweight=1 if waz<-2
replace underweight=0 if waz>=-2
replace underweight=. if waz==.
label define underweight 0"non underweight" 1"underweight",replace
label value underweight underweight
tab stunting if year==** [iweight= swt]
tab wasting if year==** [iweight= swt]
tab underweight if year==** [iweight= swt]

I have gone through again all the merging process of all the datafiles of the years together and yet I am still getting the same results or is there something I might have ommited?

Thank you in anticipation for your response

Regards,
Babs

[Updated on: Tue, 03 May 2022 12:38]

Report message to a moderator

Re: Slight difference in the Senegal Stunting, Wasting, Underweight prevalence [message #24383 is a reply to message #24371] Thu, 05 May 2022 12:37 Go to previous messageGo to next message
Janet-DHS is currently offline  Janet-DHS
Messages: 880
Registered: April 2022
Senior Member
Following is response from DHS Research & Data Analysis Director, Tom Pullum:

I see some possible problems with your code. First, the three Z scores have different legal ranges. The ranges are for the HAZ are -6 to +6, for the WAZ are -6 to +5, for the WHZ are -5 to +5. Second, and probably the main problem, is that a line such as "replace stunting=0 if haz>=-2" will include the NA (.) cases as 0. You can replace lines such as these:

gen stunting=1 if haz<-2
replace stunting=0 if haz>=-2
replace stunting=. if haz==.
With just the following following two lines:

gen stunting=0 if haz<6
replace stunting=1 if haz<-2
For any other values of the haz, stunting will be a dot, so you don't need a third line.

Stunting, etc., can be calculated from the KR file as well as the PR file. It is then limited to children who are living with the mother, but can be linked to information about the child and mother that is not in the PR file.
Re: Slight difference in the Senegal Stunting, Wasting, Underweight prevalence [message #24394 is a reply to message #24383] Sat, 07 May 2022 12:11 Go to previous messageGo to next message
tundexZ is currently offline  tundexZ
Messages: 4
Registered: March 2022
Member
Dear Tom,

Many thanks for this lines of codes. I have actually tried to use this codes based on your suggesstion but i am still getting the same stunting, wasting and underweight results.

I still look forward, perhaps if there is any other and further you can recommend based on this differrences in the results.

Regards,
Babs
Re: Slight difference in the Senegal Stunting, Wasting, Underweight prevalence [message #24400 is a reply to message #24394] Tue, 10 May 2022 10:14 Go to previous message
Janet-DHS is currently offline  Janet-DHS
Messages: 880
Registered: April 2022
Senior Member
Following is response from DHS Research & Data Analysis Director, Tom Pullum:

Are you using the sample weights? Also note that you will get slightly different estimates if you make these estimates from the PR file or the KR file. In the PR file, the anthropometry variables are hc70 hc72 and refer to all children in the household. In the KR file, they are hw70 hw72 and they refer to slightly fewer children, those in the household whose mother was eligible for the women's survey and was interviewed. If you are trying to match numbers in the report, be sure you are using the right file.

If you still do not get a match, please give me a specific number in a specific table that you cannot match. I'll try to help.

[Updated on: Tue, 10 May 2022 10:16]

Report message to a moderator

Previous Topic: MDD-W for Rwanda
Next Topic: Sierra Leone 2019
Goto Forum:
  


Current Time: Sat Nov 9 05:12:41 Coordinated Universal Time 2024