CI Decmposition Results not Showing percentage contribution

[Nthato Ras]

Hi All,

I am using Stata 16 and trying to decompose the concentration index but the results show nothing for percentage contribution. My dependent variable is a binary variable, obese or not obese. My dependent variables are quintile , feedingscheme (whether a student participates in school feeding program or not), adorace(ado for adolescent, African/White/Coloured/Asian), adogender(male/female), employ_m (whether mother is employed), educ_m and wgt_m.

The commands are as follows: using Erreygers

sca CI=r(CI)
global X  quintile feedingscheme adorace adogender empl_m educ_m wgt_m
qui sum obese [aw=wt]
sca m_obese=r(mean)
qui glm obese $X [aw=wt], family(binomial) link(logit)
qui margins , dydx(*) post

foreach x of varlist $X {
sca b_`x'=_b[`x']
}
foreach x of varlist $X {
qui{
conindex `x' [aw=wt], rankvar(quintile) truezero
sca CI_`x' = r(CI)
sum `x' [aw=wt]
sca m_`x'=r(mean)
sca elas_`x' = b_`x'*m_`x'
sca con_`x' = 4*elas_`x'*CI_`x'
sca prcnt_`x' = (con_`x'/CI)*100
}
di "`x' elasticity:", elas_`x'
di "`x' concentration index:", CI_`x'
di "`x' contribution:", con_`x'
di "`x' percentage contribution:", prcnt_`x'
}
matrix Aaa = nullmat(Aaa) \ (elas_`x', CI_`x', con_`x', prcnt_`x')
}
matrix rownames Caa= $X
matrix colnames Caa = "Elasticity""CI""Absolute""%"
matrix list Caa, format(%8.4f)
clear matrix

The results show just a fullstop next to each variable percentage contribution. And it also gives an error message "elas_ not found". I have defined elasticity

quintile elasticity: .02855949
quintile concentration index: .26154029
quintile contribution: .02987783
quintile percentage contribution: .
feedingscheme elasticity: .03727037
feedingscheme concentration index: .05987572
feedingscheme contribution: .00892636
feedingscheme percentage contribution: .
adorace elasticity: -.01348892
adorace concentration index: .01036392
adorace contribution: -.00055919
adorace percentage contribution: .
adogender elasticity: .02694672
adogender concentration index: -.00403962
adogender contribution: -.00043542
adogender percentage contribution: .
empl_m elasticity: .01834728
empl_m concentration index: .06688754
empl_m contribution: .00490882
empl_m percentage contribution: .
educ_m elasticity: .02113921
educ_m concentration index: .08112557
educ_m contribution: .00685972
educ_m percentage contribution: .
wgt_m elasticity: .15021019
wgt_m concentration index: .01183257
wgt_m contribution: .00710949
wgt_m percentage contribution: .

. matrix Aaa = nullmat(Aaa) \ (elas_`x', CI_`x', con_`x', prcnt_`x')
elas_ not found
r(111);

My dataex results are as follows:

* Example generated by -dataex-. To install: ssc install dataex
clear
input byte quintile float(feedingscheme adorace adogender empl_m educ_m wgt_m)
. . . . . . .
. . . . . . .
. . . . . . .
. . . . 1 1 2
1 1 2 2 1 1 2
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . 1 1 2
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . 2 1 3
. . . . 2 1 3
. . . . 2 1 3
. . . . . . .
2 2 2 2 2 1 3
. . . . 2 1 3
. . . . . . .
. . . . . . .
. . . . . . .
3 1 2 2 . . .
3 1 2 1 . . .
. . . . . . .
. . . . . . .
3 1 2 2 . . .
. . . . 1 1 3
. . . . . . .
. . . . . . .
3 1 2 1 2 1 2
3 1 2 2 2 1 2
. . . . 2 1 2
. . . . 2 1 2
1 1 2 1 . . .
1 1 2 1 . . .
1 . 2 1 . . .
1 1 2 1 . . .
. . . . . . .
. . . . . . .
1 . 2 2 . . .
1 1 2 2 1 1 3
. . . . . . .
. . . . . . .
. . . . . . .
1 1 2 1 1 1 3
1 1 2 2 1 1 2
1 1 2 2 1 1 3
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
1 1 2 2 . . .
. . . . . . .
. . . . . . .
1 1 2 2 . . .
1 1 2 1 . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . 1 2 2
. . . . . . .
. . . . . . .
. . . . 1 1 2
. . . . . . .
. . . . 1 2 2
2 1 2 2 1 2 2
. . . . 1 2 2
1 . 2 1 1 1 1
. . . . . . .
. . . . . . .
2 1 2 1 2 1 3
2 2 2 2 2 1 3
2 . 2 2 2 1 3
. . . . . . .
2 1 2 2 2 1 3
. . . . 1 3 3
2 1 2 2 2 1 3
. . . . 1 1 3
. . . . 1 1 3
. . . . . . .
. . . . 1 2 3
. . . . 1 2 3
. . . . 1 2 3
. . . . 1 2 3
. . . . 1 1 3
. . . . . . .
1 2 2 2 1 1 3
. . . . . . .
. . . . 1 2 3
. . . . 1 2 3
1 1 2 1 1 1 3
1 1 2 2 1 1 3
. . . . 1 1 3
. . . . 1 1 3
. . . . . . .
1 2 2 1 1 2 3
end
label values feedingscheme feedingscheme5
label def feedingscheme5 1 "fscheme", modify
label def feedingscheme5 2 "nfscheme", modify

Please assist.

Regards
Nthato

[William Lisowski]

It seems likely that the missing percentages are because the scalar CI is in fact a missing value or zero at that point, or else you also have a variable in your dataset whose name begins with CI, which may cause it to take precedence over the scalar of the same name. You should

di "CI: ", CI, scalar(CI)

immediately after it is defined, and just before the calculation of prcnt_`x', and I think you'll see that CI is not what you think it is.

The error on the matrix command occurs because that command is outside of the loop over the local macro x, so that local macro is undefined at that point, and the variable name elas_`x' becomes elas_x.

[Nthato Ras]

Hi William,

Thanks for the response.

I do not have any variable that begins with CI. I put the code below as per suggestion

di "CI: ", CI, scalar(CI)

and got an error message "scalar not found".

How do I define sca CI? I couldn't find anything on the internet.

However, my concentration index is as per below.

. conindex obese [aweight=wt], rankvar(quintile) erreygers bounded limits (0 1)

Index: No. of obs. Index value Std. error p-value
-
Erreygers norm. CI 20341 .0467093 .00407087 0.0000

[Nthato Ras]

Hi William,

I think you responded but the message vanished as I was about to click on it. Please re submit if you responded.

Thanks

[William Lisowski]

The first line in the code in post #1 is

sca CI=r(CI)

which defines the scalar CI as the concentration index returned as r(CI) by the conindex command which should be run immediately before your code in post #1.

[Nthato Ras]

It was my mistake. I did not run conindex as part of the whole loop. I ran it first, followed by the loop from sca CI.

Thank you William!

[Aklilu Sinshaw]

Hi all,
I am new for health inequity index analysis using STATA. Health variable has bivariable outcome (Yes, No) and wealth index by quantile (Poorest, poor, moderate, rich, richest). I also plan to decompose concentration index to other need and non-need categorical variables.
I would be happy if you advise me about this issue?
Thank you!

[Nthato Ras]

Dear Aklilu,

Welcome! You have come to the correct platform for advise. I am very happy with the assistance I have received from this platform. If you want to get advise quickly, please read FAQs first. This explain how to ask questions so that everyone can understand and look at you data to assist. You will also need to install ssc so that we can see your data. This increases the chance of your question being responded to quickly.

Thanks
Nthato