The update does not pop up every time for me either.  Can't tell what internal SAS EG logical rules govern the pop-up.  Seems like the program that receives the external edit needs to be open or needs to be changed internally without a save.  Otherwise, not all linked external SAS programs in an SAS EG project receive pop-up notifications.    Our SAS EG Projects contain hundreds of linked external SAS programs.  They get used in stages: create datasets, run analyses, create reports, etc.  Only a small portion of the linked external SAS Programs get used at any given time depending on the stage of the project.  It makes sense that external edits are not flagged if they are done to the programs that have not been opened in the current SAS EG session.   I would like to have a thorough documentation on the logic of the pop-up appearance.     Thanks SAS,   ... View more

Not if you use the same program.  You can open a new program window.  Or you can write two PROC FREQs and run them both.  Or you could save the first run as SAS Report and then run the second one.  Many work-arounds but SEG does not work the same way as SAS in this respect as far as I can tell.  I've been using SEG exclusively for the past decade or more and stopped looking to replicate Base SAS behavior long time ago.  You get used to it. ... View more

Unfortunately I have not used MULTINOMIAL regression since school. (LS)ESTIMATE statement should let you calculate whatever you need out of the model. Good luck! ... View more

I thought I was very specific: I want to get a proportion of patients AGE>65 out of a dataset that has a variable AGE using PROC MEANS.  I know how to do it with intermediate datasets/views and PROC SQL so I am not looking for alternative solutions.  I don't know how to be more specific. ... View more

Yes, I suppose you can look at it this way--both procs use CLASS consistently. It does help my cause, though, does it 😞 ... View more

I would not worry about banding as long as there is no systematic increase/decrease in residuals by Linear Predictor.  If your outcome is highly concentrated in some categories, you may want to analyze it as a discrete outcome rather than continuous.   Which distribution fit your data better: lognormal or Poission?  By the looks of your residual plot, you may need to look at other zero-truncated distributions as well for a good match.   Plots at Time 1 and Time 2 should be histograms.  That will enable you to see clustering.  Scatterplots are superimposed and you can't see the frequency of responses at each level. ... View more

Right, I know I can make the server do the calculations. Unfortunately, I cannot modify the datasets stored on the server. Views take an unacceptably long time to create for practical use. Adding variables such as Xgt1 = X>1 to a dataset for the purposes of calculating percentages means I need to create new interim temporary datasets ☹. Furthermore, PROC FREQ is not very friendly when it comes to creating OUTPUT datasets. My workflow relies on saving formatted output in standard tables. PROC MEANS creates beautiful OUTPUT datasets. PROC FREQ does not ☹. I hope you can appreciate the want… ... View more