The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » Ethiopia » MDD and Sample size mismatch
MDD and Sample size mismatch [message #22943] Wed, 09 June 2021 09:17 Go to next message
Hassen
Messages: 121
Registered: April 2018
Location: Ethiopia,Africa
Senior Member
Dear DHS Experts, I used Ethiopia Mini-DHS 2019 ETKR81DT dataset for analysis and Stata 14/SE software. Using standardized Stata command from Github website from chapter 11 to match the result of Minimum Dietary Diversity in figure 9.5 (MDD=14%) and table 9.7 (MDD=14.4%) and also, the number of all children age 6-23 months was 1,463. The results of my analysis using Github Stata syntax for MDD and the results in figure 9.5 & table 9.7 were mismatch. So, I need your help please? Any Stata command to match the result?
Cheers, Hassen


Hassen Ali(Chief Public Health Professional Specialist)
Re: MDD and Sample size mismatch [message #22953 is a reply to message #22943] Thu, 10 June 2021 13:38 Go to previous messageGo to next message
Shireen-DHS is currently offline  Shireen-DHS
Messages: 67
Registered: August 2020
Location: USA
Senior Member
Hello Hassen,

I was able to match the final report using the GitHub code. I think perhaps you didn't check the main file first (!NTmain.do). There are certain variables there you need to select the correct children. This was explained in the main do file. Mainly, before running the NT_IYCF.do file you need to run the following code (lines 62 to 101 in the main file) to select the correct children:

**** child's age ****
gen age = v008 - b3

* to check if survey has b19, which should be used instead to compute age.
scalar b19_included=1
capture confirm numeric variable b19, exact
if _rc>0 {
* b19 is not present
scalar b19_included=0
}
if _rc==0 {
* b19 is present; check for values
summarize b19
if r(sd)==0 | r(sd)==. {
scalar b19_included=0
}
}

if b19_included==1 {
drop age
gen age=b19
}

* Note: The following do files select for the youngest child under 2 years living with the mother. Therefore some cases will be dropped.
* Selecting for youngest child under 24 months and living with mother
keep if age < 24 & b9 == 0
* if caseid is the same as the prior case, then not the last born
keep if _n == 1 | caseid != caseid[_n-1]

Following this you can run the following to get the correct estimates for minimum dietary diversity as shown in Table 9.7:
gen wt=v005/1000000

*Among breastfeeding children
tab nt_mdd if nt_bf_curr==1 [iw=wt]
*gives 14.4 and 1246 children as in the report

*Among non-breastfeeding children
tab nt_mdd if nt_bf_curr==0 [iw=wt]
*gives 8% and 217 children as in the report

*Among all children
tab nt_mdd [iw=wt]
*gives 13.5 and 1463 children as in the report


Thank you.
Best,

Shireen Assaf
The DHS Program
Re: MDD and Sample size mismatch [message #22960 is a reply to message #22953] Fri, 11 June 2021 04:41 Go to previous message
Hassen
Messages: 121
Registered: April 2018
Location: Ethiopia,Africa
Senior Member
Dear Shireen Assaf, Fantastic Your guidance is so helpful. Thank you!!

Sincerely yours, Hassen


Hassen Ali(Chief Public Health Professional Specialist)
Previous Topic: Neonatal Mortality
Goto Forum:
  


Current Time: Sat Jun 19 17:03:08 Coordinated Universal Time 2021