[jperdue1]

Thanks Richard, Actually I was just looking at http://www.stata-journal.com/sjpdf.h...iclenum=st0137.

Best regards,
Michael

[jperdue1]

Dear Richard,
Thank you for all of your help. I did as you suggested, it works, I'm a happy camper. I just have one question. what does 'N=e(N)' do?
Best regards,

Michael

[Richard Williams]

That is the sample size for the regression. Other ereturned values could be added too, e.g. The F value. See the help for statsby.

[Richard Williams]

Also, if you want to compute p values for T statistics, the statsby command should include something like dfr = e(df_r). But if you have 1000s of cases for each regression, you can get by with just treating the Ts as Zs and calculate the P value accordingly.