simple question

[mubeen afzal]

Hi I have a simple binary logistic regression model I would like to test on an independent dataset. How do i go about testing wether it categorizes my outcome variable successfuly in the independent dataset.
Thank you

[Carlo Lazzaro]

Mubeen:
welcome to this forum.
Do you mean something along the following lines?

. use "https://www.stata-press.com/data/r17/auto.dta"
(1978 automobile data)

. logistic foreign mpg

Logistic regression                                     Number of obs =     74
                                                        LR chi2(1)    =  11.49
                                                        Prob > chi2   = 0.0007
Log likelihood = -39.28864                              Pseudo R2     = 0.1276

------------------------------------------------------------------------------
     foreign | Odds ratio   Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
         mpg |   1.173232   .0616975     3.04   0.002      1.05833    1.300608
       _cons |   .0125396   .0151891    -3.62   0.000     .0011674    .1346911
------------------------------------------------------------------------------
Note: _cons estimates baseline odds.

. predict pr_positive_outcome, p

. list foreign mpg pr_positive_outcome in 1/10

     +---------------------------+
     |  foreign   mpg   pr_pos~e |
     |---------------------------|
  1. | Domestic    22   .2964837 |
  2. | Domestic    17   .1593718 |
  3. | Domestic    22   .2964837 |
  4. | Domestic    20   .2344013 |
  5. | Domestic    15   .1210596 |
     |---------------------------|
  6. | Domestic    18   .1819565 |
  7. | Domestic    26   .4439737 |
  8. | Domestic    20   .2344013 |
  9. | Domestic    16   .1391136 |
 10. | Domestic    19   .2069539 |
     +---------------------------+

.

[mubeen afzal]

Hi carlos,
Thanks for replying. I have one continuous variable used as a predictor and one categorical binary variable. After the first command, I want to test the model on another dataset? Do I have to delete the dataset and paste the new dataset. How do I use the model from the previous (train dataset) to check if it categorizes the test dataset?

[Carlo Lazzaro]

Mubeen:
just check the very same code on the new dataset, that should be opened in a new .dta file.