How to get predicted probabilities when using logit regression

[Soffo]

~~Hi everyone. I have run a logit regression, and the output data comes in the form of odds ratio. Is there a way to transform odds ratio to predicted probabilities, so that the output will be easier to interpret? Thank you

[Maarten Buis]

You can get the predicted probabilities by typing predict pr after you have estimated your logit model. This will create a new variable called pr which will contain the predicted probabilities. After that you tabulate, and graph them in whatever way you want. However, you are probably looking the margins command.

[Soffo]

Thank you so much for your help, Maarten. Have a nice summer!

[Soffo]

Can I ask you something more, Maarten?

I did what you said. Can I then take my new pr-variable, and run it in logit, as I did whit the old variable?
I tried to do this, but it looked like the value of the Pseudo R2 became smaller in the new logit.
Is what I did OK to do, or does it give me the wrong results? Maybe it's a bad sign that the Pseudo R2 changed when I used the pr variable in stead of the original variable.

[Richard Williams]

I am not sure what you did Soffo. With logit the dependent variable is 0/1 or, more precisely, 0 / Not 0. If you tried using your computed PR variable it would have changed all the non-zero values to 1. But I wouldn't expect it to run at all since you wouldn't have any values that exactly equaled zero. In any event why would you want to use the pr variable as the dependent variable in a new logit model? Anyway, you might show your code and output and elaborate on what it is you are trying to do.

For an overview of the margins command, see

http://www3.nd.edu/~rwilliam/stats/Margins01.pdf

[Adam Mayer]

In addition to the excellent resources provided by Dr. Williams I highly recommend this book from StataPress. It provides an exhaustive overview of the margins command:

http://www.stata.com/bookstore/inter...ession-models/