If you have SAS 8 you can use the ODS features to have access to everything concern to the procedure; so any data generated by the procedure are available to the user. That's a big difference between SAS 8 and SAS 6.12 because for some procedure, especially with SAS/STAT, data, test, statistics measures available on the SAS output, they are not available in the output dataset generated by the procedure itself. So the past solution with 6.12 (or at least the one I adopted) was to re-read the SAS output with a datastep and try to find the position in the SAS output where the measure I need is.

Back to SAS 8, if you have it, you can do the following tasks I used for the proc GLM (I can't provide an example with proc LOGISTIC because in my actual company we did not yet installed SAS 8), bu you can do the same for proc LOGISTIC:

/****First run your procedure in this way (between ods trace on and ods trace off)****/ ods trace on; proc glm data=data.abpm; model pas=pad; run; ods trace off;

/****Take a look to the LOG****/ Output Added: ------------- Name: NObs Label: Number of Observations Template: Stat.GLM.NObs Path: GLM.Data.NObs -------------

Output Added: ------------- Name: OverallANOVA Label: Overall ANOVA Template: stat.GLM.OverallANOVA Path: GLM.ANOVA.pas.OverallANOVA -------------

Output Added: ------------- Name: FitStatistics Label: Fit Statistics Template: stat.GLM.FitStatistics Path: GLM.ANOVA.pas.FitStatistics -------------

Output Added: ------------- Name: ModelANOVA Label: Type I Model ANOVA Template: stat.GLM.Tests Path: GLM.ANOVA.pas.ModelANOVA -------------

/****Identify the portion of the procedure output you need (path), then re-run the proc as follows****/ ods output GLM.ANOVA.pas.FitStatistics=MyTest1 GLM.ANOVA.pas.OverallANOVA=MyTest2; proc glm data=data.abpm; model pas=pad; run;

In my case, Datasets MyTest1 and dataset MyTest2 will contain respictively output measure concerning Statistics Fitted and Overall Anova.

Hope this will fit to your question Bye Angelo

Tinazzi Angelo Sebastiano Lead Analyst Dept. of Statistics and Programming (Bld. 31) Pharmacia Italy Viale Pasteur, 10 I - 20014 Nerviano (Milano) +39 02 4838 5747

>-----Original Message----- >From: Terry Fudin [mailto:t.fudin@GENAISSANCE.COM] >Sent: Wednesday, January 30, 2002 7:59 PM >To: SAS-L@LISTSERV.UGA.EDU >Subject: PROC LOGISITIC > > >I am trying to output the p-values (prob) from a proc logistic into a sas >dataset. When I use outest in the proc statement, it only gives the >parameter estimates. When I use the output statement, it creates a dataset >with all of the variables on the input dataset and their predictors, >standard errors, and confidence limits, etc. However I cannot find a way >to generate a dataset with the Wald chi-square p-values for the effects. I >also cant find a way to output the odds ratios and confidence limits for >the effects. Is there a way to do this?