Conditional Logistic Regression - Formula used in Predict command

[Hans Schubert]

Good evening,

Would anyone know the formula used in the predict command for a conditional logistic regression (final_prob) column in this example.

Thank you

* Example generated by -dataex-. For more info, type help dataex
clear
input double caseid float response double jsh_321 float final_prob
674335 48.6985   .5  .3067266
674339 40.5106 -1.4 .17097665
674337 37.2532 -.74 .15330617
674340 36.6393   .8    .16991
674334 25.0731  3.1 .11490745
674336 23.5891   .4  .0841731
end

Conditional (fixed-effects) logistic regression Number of obs = 644,019
LR chi2(2) = 39026.19
Prob > chi2 = 0.0000
Log likelihood = -128903.49 Pseudo R2 = 0.1315

------------------------------------------------------------------------------
won | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
response | .0511508 .0002847 179.67 0.000 .0505928 .0517087
jsh_321 | .0871644 .0030459 28.62 0.000 .0811945 .0931344
------------------------------------------------------------------------------

[William Lisowski]

A more helpful statement of your question would have included the actual commands you ran and the output from those commands.

So the best we can do is give generic advice. Run

help clogit postestimation

Scroll down and click on predict

Reading that, you will see that the default prediction is

pc1, the default, calculates the probability of a positive outcome conditional on one positive outcome within group.

Then, scroll back up to the top of the window, and click on Methods and formulas

Acrobat Reader will open to the appropriate page in the full PDF documentation included in your Stata installation and accessible through Stata's Help menu. There you can see in all its glory the for pc1.