Format Table: Average Marginal Effects of Multinomial Logistic Regression

[Melissa Garcia]

I would like to create a table of average marginals effects estimated from a multinomial logistic regression model. I would like each outcome category (general, academic, vocation) to have its own column and SEs to be listed below the coefficients.

The below mtable code from Spot13 works.

use https://stats.idre.ucla.edu/stat/data/hsbdemo, clear
mlogit prog i.female i.ses read write 
mtable, dydx (female) rowname(female) statistics (est se) 
 foreach var of varlist ses read write {
               mtable, dydx (`var') rowname (`var') statistics (est se) below 
        }

But is there another possible solution using margins, dydx(*) and estout (from SSC). Given the number of tables and variables I am working with, I was wondering if there's an easier solution than the one I found.

[Andrew Musau]

use https://stats.idre.ucla.edu/stat/data/hsbdemo, clear
estimates clear
eststo mlogit: mlogit prog i.female i.ses read write 
levelsof `e(depvar)', local(levels)
local mlab
foreach lev of local levels{
    est restore mlogit
    local mlab `mlab' "`:lab (`e(depvar)') `lev''"
    margins, dydx(*) predict(pr outcome(`lev')) post
    est sto est`lev'
} 
esttab est*, se nobaselevels mlab(`mlab') star(* 0.1 ** 0.05 *** 0.01)

Res.:

. esttab est*, se nobaselevels mlab(`mlab') star(* 0.1 ** 0.05 *** 0.01)

------------------------------------------------------------
                      (1)             (2)             (3)   
                  general        academic        vocation   
------------------------------------------------------------
1.female          -0.0404         -0.0139          0.0543   
                 (0.0639)        (0.0686)        (0.0599)   

2.ses              -0.121         0.00578           0.115*  
                 (0.0801)        (0.0829)        (0.0659)   

3.ses              -0.147*          0.175*        -0.0278   
                 (0.0893)        (0.0954)        (0.0749)   

read             -0.00349          0.0113***     -0.00784** 
                (0.00374)       (0.00391)       (0.00374)   

write            0.000339          0.0104**       -0.0107***
                (0.00392)       (0.00419)       (0.00358)   
------------------------------------------------------------
N                     200             200             200   
------------------------------------------------------------
Standard errors in parentheses
* p<0.1, ** p<0.05, *** p<0.01

[Melissa Garcia]

Thank you, Andrew. This was very helpful!