Multicolumn output of stargazer to be used in knitr

Question

Here is MWE:

library(pscl)

data("bioChemists", package = "pscl")

fm_pois <- glm(art ~ ., data = bioChemists, family = poisson)
fm_qpois <- glm(art ~ ., data = bioChemists, family = quasipoisson)
fm_nb <- glm.nb(art ~ ., data = bioChemists)
fm_zinb <- zeroinfl(art ~ . | 1, data = bioChemists, dist = "negbin")

library(stargazer)
stargazer(
  fm_pois, fm_qpois,  fm_nb, fm_zinb
  , type = "text"
)

=============================================================================
                                      Dependent variable:                    
                  -----------------------------------------------------------
                                              art                            
                   Poisson   glm: quasipoisson     negative     zero-inflated
                                link = log         binomial      count data  
                     (1)            (2)              (3)             (4)     
-----------------------------------------------------------------------------
femWomen          -0.225***      -0.225***        -0.216***       -0.216***  
                   (0.055)        (0.074)          (0.073)         (0.073)   

marMarried         0.155**        0.155*            0.150*         0.150*    
                   (0.061)        (0.083)          (0.082)         (0.082)   

kid5              -0.185***      -0.185***        -0.176***       -0.176***  
                   (0.040)        (0.054)          (0.053)         (0.053)   

phd                 0.013          0.013            0.015           0.015    
                   (0.026)        (0.036)          (0.036)         (0.036)   

ment               0.026***      0.026***          0.029***       0.029***   
                   (0.002)        (0.003)          (0.003)         (0.003)   

Constant           0.305***       0.305**           0.256*         0.256*    
                   (0.103)        (0.139)          (0.137)         (0.139)   

-----------------------------------------------------------------------------
Observations         915            915              915             915     
Log Likelihood    -1,651.056                      -1,561.958     -1,560.959  
theta                                          2.264*** (0.271)              
Akaike Inf. Crit. 3,314.113                       3,135.917                  
=============================================================================
Note:                                             *p<0.1; **p<0.05; ***p<0.01

I'm looking for multicolumn output like this:

=============================================================================
                                      Dependent variable:                    
                  -----------------------------------------------------------
                                              art                            
                           Poisson              Negative Binomial 

                    Poisson     QuasiPoisson         NB           ZINB

                     (1)            (2)              (3)             (4)     
-----------------------------------------------------------------------------
femWomen          -0.225***      -0.225***        -0.216***       -0.216***  
                   (0.055)        (0.074)          (0.073)         (0.073)   

marMarried         0.155**        0.155*            0.150*         0.150*    
                   (0.061)        (0.083)          (0.082)         (0.082)   

kid5              -0.185***      -0.185***        -0.176***       -0.176***  
                   (0.040)        (0.054)          (0.053)         (0.053)   

phd                 0.013          0.013            0.015           0.015    
                   (0.026)        (0.036)          (0.036)         (0.036)   

ment               0.026***      0.026***          0.029***       0.029***   
                   (0.002)        (0.003)          (0.003)         (0.003)   

Constant           0.305***       0.305**           0.256*         0.256*    
                   (0.103)        (0.139)          (0.137)         (0.139)   

-----------------------------------------------------------------------------
Observations         915            915              915             915     
Log Likelihood    -1,651.056                      -1,561.958     -1,560.959  
theta                                          2.264*** (0.271)              
Akaike Inf. Crit. 3,314.113                       3,135.917                  
=============================================================================
Note:                                             *p<0.1; **p<0.05; ***p<0.01

1. First row should have the word Poisson for first two columns and Negative Binomial for next two columns.
2. Second row should have columns names like Poisson, Quasi Poisson, Negative Binomial and Zero Inflated Negative Binomial.

I found this link but could not figured out how to get this one.

Gass

I don't think this can be done either using stargazer or texreg. You'd need to modify the source to produce your desired output. Unless this is something you're doing all the time, it's probably much easier just to edit the output manually. – Sentinel

Answer 1

Like Nick Kennedy I do not think that stargazer can produce your desired output directly.

Therefore, here a workaround: Save the stargazer table in an object and add the desired lines manually. I hardcoded this here; with some more effort it should be possible to center the text above the respective columns automatically. Note that I slightly changed your stargazer call in order to hide the (wrong) model names.

library(pscl)
library(stargazer)

data("bioChemists", package = "pscl")

fm_pois <- glm(art ~ ., data = bioChemists, family = poisson)
fm_qpois <- glm(art ~ ., data = bioChemists, family = quasipoisson)
fm_nb <- glm.nb(art ~ ., data = bioChemists)
fm_zinb <- zeroinfl(art ~ . | 1, data = bioChemists, dist = "negbin")

byLine <- 
  do.call("c", 
          strsplit(
            capture.output(
              stargazer(fm_pois, fm_qpois,  fm_nb, fm_zinb, 
                        type = "text", model.names = FALSE)
              ),
            "\n"))

result <- append(
  byLine, 
  c(
  "                        Poisson               Negative Binomial",
  "",
  "                   Poisson  QuasiPoisson      NB          ZINB"
  ), 
  after = c(4, 5, 6))

cat(paste(result, collapse = "\n"))

# ==================================================================
#                                 Dependent variable:               
#                   ------------------------------------------------
#                                         art                       
#                         Poisson               Negative Binomial
# 
#                    Poisson  QuasiPoisson      NB          ZINB
#                      (1)        (2)          (3)           (4)    
# ------------------------------------------------------------------
# femWomen          -0.225***  -0.225***    -0.216***     -0.216*** 
#                    (0.055)    (0.074)      (0.073)       (0.073)  
#                                                                   
# marMarried         0.155**    0.155*        0.150*        0.150*  
#                    (0.061)    (0.083)      (0.082)       (0.082)  
#                                                                   
# kid5              -0.185***  -0.185***    -0.176***     -0.176*** 
#                    (0.040)    (0.054)      (0.053)       (0.053)  
#                                                                   
# phd                 0.013      0.013        0.015         0.015   
#                    (0.026)    (0.036)      (0.036)       (0.036)  
#                                                                   
# ment               0.026***  0.026***      0.029***      0.029*** 
#                    (0.002)    (0.003)      (0.003)       (0.003)  
#                                                                   
# Constant           0.305***   0.305**       0.256*        0.256*  
#                    (0.103)    (0.139)      (0.137)       (0.139)  
#                                                                   
# ------------------------------------------------------------------
# Observations         915        915          915           915    
# Log Likelihood    -1,651.056              -1,561.958    -1,560.959
# theta                                  2.264*** (0.271)           
# Akaike Inf. Crit. 3,314.113               3,135.917               
# ==================================================================
# Note:                                  *p<0.1; **p<0.05; ***p<0.01

Tweeze