Trying to understand logit tables

[Bill MacIntosh]

Hi all,

I'm fairly new to this so please bear with me.

I've run a logit model, and in it I have included the variable 'Brexit' and the variable 'political_knowledge'. The variables and coefficients themselves are not too important, but I am trying to understand why only half the values show up.

I have pasted the table below (apologies, I haven't got my head around using dataex for logit tables), and as you can see only a value for remain appears, and when I interacted brexit with political_knowledge it still only shows the remain values again.

why is this?


brexit
Remain | -1.143769 .4453552 -2.57 0.010 -2.01665 -.2708892


political_knowledge##brexit

OK#Remain | 1.376699 .6378883 2.16 0.031 .1264606 2.626937
Good#Remain | .7214613 .5805425 1.24 0.214 -.4163811 1.859304
Great#Remain | 1.188758 .5590191 2.13 0.033 .0931012 2.284416

[Nick Cox]

Presumably Remain is an indicator variable with values (0, 1). (Other names are binary, dichotomous, dummy, quantal, zero-one, Boolean.)

Whenever such a variable is a predictor (covariate, independent variable, whatever) only one coefficient is shown, namely by default the effect of that predictor being 1 not 0. (You can flip the default by changing the factor variable notation you used on your unstated logit command.)

Here's an example every Stata user can run.

. webuse nlswork, clear
(National Longitudinal Survey of Young Women, 14-24 years old in 1968)


. logit nev_mar i.collgrad i.c_city

Iteration 0:   log likelihood = -15362.312  
Iteration 1:   log likelihood = -15076.255  
Iteration 2:   log likelihood = -15073.457  
Iteration 3:   log likelihood = -15073.457  

Logistic regression                                     Number of obs = 28,510
                                                        LR chi2(2)    = 577.71
                                                        Prob > chi2   = 0.0000
Log likelihood = -15073.457                             Pseudo R2     = 0.0188

------------------------------------------------------------------------------
     nev_mar | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
  1.collgrad |   .1196052   .0372292     3.21   0.001     .0466373    .1925731
    1.c_city |   .6840089   .0286618    23.86   0.000     .6278329    .7401849
       _cons |  -1.504111   .0201021   -74.82   0.000     -1.54351   -1.464711
------------------------------------------------------------------------------


. logit nev_mar ib1.collgrad ib1.c_city

Iteration 0:   log likelihood = -15362.312  
Iteration 1:   log likelihood = -15076.255  
Iteration 2:   log likelihood = -15073.457  
Iteration 3:   log likelihood = -15073.457  

Logistic regression                                     Number of obs = 28,510
                                                        LR chi2(2)    = 577.71
                                                        Prob > chi2   = 0.0000
Log likelihood = -15073.457                             Pseudo R2     = 0.0188

------------------------------------------------------------------------------
     nev_mar | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
  0.collgrad |  -.1196052   .0372292    -3.21   0.001    -.1925731   -.0466373
    0.c_city |  -.6840089   .0286618   -23.86   0.000    -.7401849   -.6278329
       _cons |  -.7004968   .0372216   -18.82   0.000    -.7734497   -.6275439
------------------------------------------------------------------------------

.

[Bill MacIntosh]

Thank you very much Nick