Thanks Paige. There are two concerns that brought me to seek formats in PROC MEANS. The datasets I am working with are very large. Adding computed variables substantially increases the dataset size. Second, the data resides on a server to which we have relatively slow access. Moving data from the source to the locally installed SAS can take time. I suppose there is also a third consideration. Our workflow is largely based on using PROC MEANS to create summaries by hospital. I can get N, nMiss, Mean, Median, Variance, etc. but not proportion of values above a threshold. Maybe I am atypical in the kind of data summary I seek, but it seems to me that proportion of values in a variable is a common descriptive statistic people seek. I think it is reasonable to desire that a function to extract a proportion of value/values was available in PROC MEANS. It would be nice to get all the summary statistics from one PROC (incidentally, the predecessor of PROC MEANS was PROC SUMMARY). Just makes things more compact, standard, and clean. ... View more

Tamara,    LogNormal distribution cannot accept zero as the log of zero is undefined.  You need to add a constant to your outcome and remember that you did that when interpreting the results; e.g. CBI_Total+1.  That will retain the zeros.   Your residual distribution does not look dependent on the linear predictor.  I suspect that the banding is the property of your outcome not the model.  Plot residuals with CBI_Total on the X-axis.  What do you see? Your residual distribution looks quite normal except for a spike on the lower end.  Looks like you have an atypically common response or something like that.  Take a look at the distribution of your outcome score by Time and see if you can spot what the deviations from normality are.  If it is an atypically common single Total score, what is that and why, what does it tell you about the instrument or the population?   ... View more

The plot you are using seems to be a product of your choice of distribution more than anything else. Poisson would not be my choice for a continuous variable that you have. Have you tried lognormal? If you find that variance is not equal in your two groups, you can add a 'GROUP=GS' option to your RANDOM statement to allow for the variance estimates to be different between the two groups. Second, you are not modeling repeated measures by time. I would use a random r-side effect 'RANDOM Time / sub=ID residual type=AR(1)'. Are you sure you need TYPE=UN for your g-side matrices? Most commonly, measurements for different individuals in a study are not correlated and are modeled as TYPE=VC. ... View more

No, not exactly, SASKiwi. What I want is for PROC MEANS to average the discrete values of Age 0 and 1 rather than actual age values. Say there are only two ages 60 and 70. PROC MEANS would return a mean of 65 for these two. What I want it to do is apply the format to these values, turn them into 0 and 1 (because 60 is less than 65 and 70 is more) and return the average of 0.50.  If there are four ages: 40, 50, 60, and 70, they should be formatted as 0, 0, 0, and 1 and the mean I want to get is 0.25 rather than 55. PROC SQL does what I need if I code it as: proc sql; select MEAN(AgeCHDdiag>65) from SASHelp.Heart; quit; ... View more

Bummer 😞 Why would PROC MEANS be different from PROC FREQ? I wonder if PROC REPORT of PROC TABULATE would offer an option to use the formatted values? ... View more

Thanks for an additional way to accomplish the task. Now, can you do this in SAS without the interim DATA step? ... View more

First, your output is not BINOMIAL but rather MULTINOMIAL. Second, you do not have any r-side effects to account for repeated measures. Your code should look something like this: proc glimmix data=out2.Client_VL_Euci44_f; class <SubjectID> euci44 year raceeth; model vl_status =year raceeth /solution dist=multinomial link=logit; random Year / sub=<SubjectID> _residual_ type=ar(1); lsmeans raceeth / om ilink; format vl_status vlstatusfmt. gender genderfmt. raceeth raceethfmt.; run; ... View more

Generically speaking, SAS takes care of the imbalance for you. But, if you give us the research question you are trying to test and more details on your design you may get useful input. For example, do you know that Diagnosis is not used in the model structure of this GLIMMIX model outside of distinguishing among IDs across different diagnoses? Your model has a g-matrix with N*N dimensions. Diagnosis is not part of your model and, therefore, your balancing question is not relevant. ... View more

I would like to help but I have no idea what this question is asking. Can you be more detailed? ... View more

@Patrick , I get the 'copy', 'import' and 'open' parts.  That's not a problem.  I don't get why you are completely sidestepping the 'embed' part.  Are you saying I don't need to do that to accomplish my goal?  The key requirement is for four different programs to 'feed' off of the common code making small program-specific adjustments; e.g., 'Age GE 16' for one program and 'Age LE 18' for another, without sending those small changes to the mothership on the network.  If I don't embed, I need to make four copies of the source programs every cycle and that is exactly what I am trying to avoid.  EMBED does exactly what I need--it breaks the link between the program on the network and stores all changes in SEG project.  It's kinda like a coloring book, I want to import the outlines but each user gets to color in their own unique way.   For versioning, that's exactly what we do: simply copy all of the code from Fall cycle to Spring cycle to Fall cycle and so on.  The small variations between the programs we mostly manage with program-specific macro variables.  The ad hoc ones get saved in the logs. ... View more