Adding group name to categorical variables in esttab command

[Shir Hefer]

Hi,
I am trying to add a heading for each categorical variable in my regression.
This is the code I am now using:

oprobit life_satisfaction population ${x}
est store est1

//LS for jews:
oprobit life_satisfaction ${x} i.migration i.migrate_year if population==1
est store est2

//LS for arabs:
oprobit life_satisfaction ${x} i.religion_arab
est store est3

esttab est1 est2 est3 using ls.doc, replace noomitted nobaselevels se pr2 label mlabels("Pooled" "Jews" "Arabs")


To generate the outcome:


--------------------------------------------------------------------
(1) (2) (3)
Pooled Jews Arabs
--------------------------------------------------------------------


married 0.251*** 0.243*** 0.245
(0.0558) (0.0645) (0.127)

divorced/separated -0.0670 -0.0194 -0.236
(0.0755) (0.0816) (0.254)

widow(er) 0.0453 0.0218 0.236
(0.0966) (0.106) (0.261)


1-<2 -0.0692 -0.0755 -0.0791
(0.0434) (0.0475) (0.120)

2+ -0.255** -0.127 -0.301
(0.0839) (0.115) (0.174)



I would like the outcome to be similar to the one below: (now added it by hand, but am searching for a code....)

--------------------------------------------------------------------
(1) (2) (3)
Pooled Jews Arabs
--------------------------------------------------------------------

Marital_status

married 0.251*** 0.243*** 0.245
(0.0558) (0.0645) (0.127)

divorced/separated -0.0670 -0.0194 -0.236
(0.0755) (0.0816) (0.254)

widow(er) 0.0453 0.0218 0.236
(0.0966) (0.106) (0.261)
Room_pp

1-<2 -0.0692 -0.0755 -0.0791
(0.0434) (0.0475) (0.120)

2+ -0.255** -0.127 -0.301


For the example above, I would like to add "marital_status" above the categorical variable and "room_pp" above another categorical variable. These are the names of the variables in my dataset
I have been looking up statalist but didn't find a simple solution for this.
Hope my answer is clear enough.

Thank you,
Shir.

[Andrew Musau]

estout/ esttab is from Stata Journal (FAQ Advice #12). See the -refcat- option of estout.

webuse lbw
probit low age i.race smoke
esttab, refcat(2.race "Mother's race", nolab) lab nobaselevels

Res.:

. esttab, refcat(2.race "Mother's race", nolab) lab nobaselevels

------------------------------------
                              (1)   
                     birthw~2500g   
------------------------------------
birthweight<2500g                   
age of mother             -0.0219   
                          (-1.09)   

Mother's race                       

black                       0.618*  
                           (2.06)   

other                       0.646** 
                           (2.69)   

smoked during preg~y        0.676** 
                           (3.06)   

Constant                   -0.606   
                          (-1.17)   
------------------------------------
Observations                  189   
------------------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001

[Alyona Ivashkina]

Hello! I am also facing the same need of the title for categorical variables in the regression output created in estout (or other package). However, I often have many and different factor variables, so I wonder whether there is an automated way to do this? E.g. if there is a variable label defined for the categorical variable, can it be used by the estout (or other package) to create a heading?

[Andrew Musau]

I do not believe that such an option exists. You will probably need to write a program that does that if it is something that you often require.