Negative Binomial Regression output question with IRR

[Dave Baik]

I currently have the following codes to export negative binomial regression results to excel:

nbreg $reg1 estimates store m1, title(DV=KEYWORDS)
nbreg $reg1, irr estimates store m2, title(DV=KEYWORDSPOST)
estout m1 m2, cells(b(star fmt(3)) se(par fmt(2)))
legend label varlabels(_cons Constant)


However, when i run these codes, I am only getting the m1 output with no incidence ratio coefficients. I was wondering if there is a way to export one column showing non-irr results and another column next to each other showing irr results.

Thank you,

[Andrew Musau]

Welcome to Statalist. Please take the time to review the FAQs to improve your future posts. Here are some missing details:

1. estout is from Stata Journal / SSC.
2. Incidence-rate ratios here refer to exponentiated negative binomial coefficients

With more current versions of estout, you can output the same model multiple times. The -irr- option in nbreg only affects the displayed results, i.e., it does not change the coefficients and variance matrices. However, estout has options to output exponentiated coefficients. Here is a reproducible example:

webuse rod93
generate logexp=ln(exposure)
qui nbreg deaths i.cohort, exposure(exp)
estimates store m1
esttab m1 m1, transform(exp(@), pattern(0 1))
*COMPARE WITH STATA RESULTS
nbreg deaths i.cohort, exposure(exp) nolog
nbreg deaths i.cohort, exposure(exp) nolog irr

Res.:

. esttab m1 m1, transform(exp(@), pattern(0 1))

--------------------------------------------
                      (1)             (2)  
                   deaths          deaths  
--------------------------------------------
deaths                                      
1.cohort                0               1  
                      (.)             (.)  

2.cohort           -0.268           0.765  
                  (-0.37)         (-0.37)  

3.cohort           -0.457           0.633  
                  (-0.63)         (-0.63)  

_cons              -2.087***        0.124***
                  (-4.08)         (-4.08)  
--------------------------------------------
/                                          
lnalpha             0.594*          1.811*  
                   (2.30)          (2.30)  
--------------------------------------------
N                      21              21  
--------------------------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001

.
. *COMPARE WITH STATA RESULTS

.
. nbreg deaths i.cohort, exposure(exp) nolog

Negative binomial regression                    Number of obs     =         21
                                                LR chi2(2)        =       0.40
Dispersion     = mean                           Prob > chi2       =     0.8171
Log likelihood =  -131.3799                     Pseudo R2         =     0.0015

------------------------------------------------------------------------------
      deaths |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
      cohort |
  1960-1967  |  -.2676187   .7237203    -0.37   0.712    -1.686085    1.150847
  1968-1976  |  -.4573957   .7236651    -0.63   0.527    -1.875753    .9609618
             |
       _cons |  -2.086731   .5118559    -4.08   0.000     -3.08995   -1.083511
ln(exposure) |          1  (exposure)
-------------+----------------------------------------------------------------
    /lnalpha |   .5939963   .2583615                       .087617    1.100376
-------------+----------------------------------------------------------------
       alpha |   1.811212   .4679475                       1.09157    3.005294
------------------------------------------------------------------------------
LR test of alpha=0: chibar2(01) = 4056.27              Prob >= chibar2 = 0.000

.
. nbreg deaths i.cohort, exposure(exp) nolog irr

Negative binomial regression                    Number of obs     =         21
                                                LR chi2(2)        =       0.40
Dispersion     = mean                           Prob > chi2       =     0.8171
Log likelihood =  -131.3799                     Pseudo R2         =     0.0015

------------------------------------------------------------------------------
      deaths |        IRR   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
      cohort |
  1960-1967  |   .7651995   .5537904    -0.37   0.712     .1852434    3.160869
  1968-1976  |   .6329298   .4580292    -0.63   0.527     .1532395    2.614209
             |
       _cons |   .1240922   .0635173    -4.08   0.000     .0455042    .3384052
ln(exposure) |          1  (exposure)
-------------+----------------------------------------------------------------
    /lnalpha |   .5939963   .2583615                       .087617    1.100376
-------------+----------------------------------------------------------------
       alpha |   1.811212   .4679475                       1.09157    3.005294
------------------------------------------------------------------------------
Note: Estimates are transformed only in the first equation.
Note: _cons estimates baseline incidence rate.
LR test of alpha=0: chibar2(01) = 4056.27              Prob >= chibar2 = 0.000

.