Missing categories in regression output

[Joe Tuckles]

Can anyone advise why this may be or how to rectify it?

 logit case i.tv i.comp i.cat_sleep_duration

Iteration 0:   log likelihood = -339.01006  
Iteration 1:   log likelihood = -334.28335  
Iteration 2:   log likelihood = -334.19161  
Iteration 3:   log likelihood = -334.19158  

Logistic regression                             Number of obs     =        756
                                                LR chi2(9)        =       9.64
                                                Prob > chi2       =     0.3807
Log likelihood = -334.19158                     Pseudo R2         =     0.0142

-------------------------------------------------------------------------------------------------------
                                 case |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------------------------+----------------------------------------------------------------
                                   tv |
                                   2  |  -.2806706   .2776516    -1.01   0.312    -.8248577    .2635165
                                   3  |  -.3937416   .3052837    -1.29   0.197    -.9920866    .2046035
                                   4  |   .1992853   .3122241     0.64   0.523    -.4126627    .8112333
                                      |
                             comp |
                                   2  |  -.1800197   .2925374    -0.62   0.538    -.7533825    .3933431
                                   3  |   .2965882   .2883551     1.03   0.304    -.2685775    .8617538
                                   4  |   .5721952   .4642169     1.23   0.218    -.3376533    1.482044
                                       |
                   cat_sleep_duration |
                                   2  |  -.1782777   .4567095    -0.39   0.696    -1.073412    .7168565
                                   3  |  -.3019667   .4335158    -0.70   0.486    -1.151642    .5477086
                                   4  |  -.3858412    .458861    -0.84   0.400    -1.285192    .5135099
                                      |
                                _cons |  -1.209932   .4582096    -2.64   0.008    -2.108007   -.3118582
-------------------------------------------------------------------------------------------------------

[Nick Cox]

Each factor variable with k categories implies that k - 1 coefficients will be reported, just as you get one coefficient for a indicator coding whether some binary state is true or false.

Utterly expected -- see your regression or econometrics text. Nothing to rectify unless you prefer different base levels.

[Carlo Lazzaro]

Joe:
this way Stata shelters you from the dummy trap, that is always hunting for new goals (https://en.wikipedia.org/wiki/Dummy_...le_(statistics)).