Dear members of the list,
After a multinomial logit model, I am interested in extracting the predicted probabilities corresponding to one my key independent variables for the categories of another independent variable; that is, in the predicted probabilities corresponding to an interaction effect between two independent variables in my model. I am particularly interested in these predicted probabilities for one of the possible outcomes in the multinomial logit model.
First, I use estimates store....
...and then esttab to extract the predicted probabilities to an csv file
I have simplified my path and my code, but the result in Stata is something as follows:
My apologies for the length of the output. My question comes next.
How could I manage to take these predicted probabilities to Excel so that each category in one of the independent variables in the interaction (say, field of studies, which is the first one) is placed in one column... and the other variable (parental occupation) in the rows? This would greatly help me in arranging my tables, but, in spite of my search in Statalist and in the esttab help menu (especially layout option), I do not manage to get the answer.
Thanks for your attention
And kind regards
Luis Ortiz
After a multinomial logit model, I am interested in extracting the predicted probabilities corresponding to one my key independent variables for the categories of another independent variable; that is, in the predicted probabilities corresponding to an interaction effect between two independent variables in my model. I am particularly interested in these predicted probabilities for one of the possible outcomes in the multinomial logit model.
Code:
margins var.indep1#var.indep2 , predict(outcome(Work)) post
Code:
estimates store ocup_fos_work
Code:
. esttab using myfile.csv, se label nolines replace
HTML Code:
. margins codi_a2#ocu_parents if codi_a2<6, predict(outcome(Work)) post
Predictive margins Number of obs = 33,312
Model VCE: OIM
Expression: Pr(expect3==Work), predict(outcome(Work))
-----------------------------------------------------------------------------------------------------
| Delta-method
| Margin std. err. z P>|z| [95% conf. interval]
------------------------------------+----------------------------------------------------------------
codi_a2#ocu_parents |
Humanities#Elementary/never work | .1231373 .0232878 5.29 0.000 .077494 .1687806
Humanities#Unskilled | .1767854 .0434439 4.07 0.000 .0916369 .2619338
Humanities#Skilled | .105484 .0273348 3.86 0.000 .0519087 .1590593
Humanities#Small employers | .0972449 .0190566 5.10 0.000 .0598946 .1345952
Humanities#Service | .0938936 .0133487 7.03 0.000 .0677305 .1200566
Humanities#Clerks | .1279894 .0160328 7.98 0.000 .0965657 .1594131
Humanities#Ass.Profess | .1443823 .0212913 6.78 0.000 .1026521 .1861125
Humanities#Profess Managers | .0963129 .0100046 9.63 0.000 .0767041 .1159216
Social Sc#Elementary/never work | .2325878 .0129943 17.90 0.000 .2071195 .2580561
Social Sc#Unskilled | .2233837 .0201639 11.08 0.000 .1838631 .2629042
Social Sc#Skilled | .2458289 .01904 12.91 0.000 .2085112 .2831467
Social Sc#Small employers | .2170858 .0122917 17.66 0.000 .1929946 .241177
Social Sc#Service | .2187633 .0089836 24.35 0.000 .2011558 .2363709
Social Sc#Clerks | .1917011 .0092521 20.72 0.000 .1735674 .2098348
Social Sc#Ass.Profess | .1906258 .0106476 17.90 0.000 .1697569 .2114947
Social Sc#Profess Managers | .1667551 .0056803 29.36 0.000 .155622 .1778882
Nat Sciences#Elementary/never work | .1091409 .0235161 4.64 0.000 .0630502 .1552317
Nat Sciences#Unskilled | .1619883 .0424369 3.82 0.000 .0788135 .2451631
Nat Sciences#Skilled | .1654063 .0398721 4.15 0.000 .0872583 .2435543
Nat Sciences#Small employers | .1187532 .0208774 5.69 0.000 .0778343 .1596721
Nat Sciences#Service | .1110627 .0151574 7.33 0.000 .0813548 .1407706
Nat Sciences#Clerks | .1024714 .0129927 7.89 0.000 .0770063 .1279366
Nat Sciences#Ass.Profess | .0909274 .0152161 5.98 0.000 .0611044 .1207504
Nat Sciences#Profess Managers | .0829258 .0075252 11.02 0.000 .0681767 .097675
Health Sc.#Elementary/never work | .2113802 .0209426 10.09 0.000 .1703334 .2524269
Health Sc.#Unskilled | .2328505 .0323976 7.19 0.000 .1693523 .2963486
Health Sc.#Skilled | .2340141 .0272348 8.59 0.000 .180635 .2873933
Health Sc.#Small employers | .1991124 .0168673 11.80 0.000 .1660532 .2321716
Health Sc.#Service | .2110933 .0136454 15.47 0.000 .1843489 .2378377
Health Sc.#Clerks | .2095122 .0129883 16.13 0.000 .1840556 .2349688
Health Sc.#Ass.Profess | .1901181 .014732 12.91 0.000 .1612439 .2189924
Health Sc.#Profess Managers | .2137041 .0072126 29.63 0.000 .1995678 .2278404
Engin. Arqui#Elementary/never work | .2389631 .0200133 11.94 0.000 .1997378 .2781885
Engin. Arqui#Unskilled | .2529068 .030537 8.28 0.000 .1930555 .3127582
Engin. Arqui#Skilled | .2764256 .0265191 10.42 0.000 .224449 .3284021
Engin. Arqui#Small employers | .2200667 .0170406 12.91 0.000 .1866678 .2534657
Engin. Arqui#Service | .2438614 .0138371 17.62 0.000 .2167413 .2709816
Engin. Arqui#Clerks | .2046769 .011521 17.77 0.000 .1820961 .2272576
Engin. Arqui#Ass.Profess | .1846245 .0129246 14.28 0.000 .1592927 .2099563
Engin. Arqui#Profess Managers | .1503262 .0063353 23.73 0.000 .1379092 .1627431
-----------------------------------------------------------------------------------------------------
How could I manage to take these predicted probabilities to Excel so that each category in one of the independent variables in the interaction (say, field of studies, which is the first one) is placed in one column... and the other variable (parental occupation) in the rows? This would greatly help me in arranging my tables, but, in spite of my search in Statalist and in the esttab help menu (especially layout option), I do not manage to get the answer.
Thanks for your attention
And kind regards
Luis Ortiz
