I see, that makes sense. Thank you so much for your helpful response! ... View more

@YYK273  Please start  your own thread. You can post a link to this question if you think it helpful.   Post your LOG with the code and the notes and messages as well as the error. Often there is diagnostic information that explains why. Copy the text from the log; on the forum open a text box and paste the text with the messages.   That sort of message where it includes something like in _Imputation_= 1 in the input PARMS= data set. typically indicates a data related issue related to the value of the variable, such as too few observations to model something.   ... View more

You could also use the NAMELEN= option on the PHREG statement to keep it from truncating the names of the effects. ... View more

If you are going to use the MCMC statement then you will need to create dummy variables for the categorical variables.  Unless you have a specific reason for doing so, you could avoid having to impute in two steps by simply using the FCS statement.  This would enable you to use the CLASS statement and avoid having to dummy code. ... View more

I found another paper explains with B is. It is the covariance matrix of the estimates Qˆ m. Thank you so much for your guidance. ... View more

You could use CONTRAST statements to test the nested models.  There are some good examples in the usage note linked below. 24447 - Examples of writing CONTRAST and ESTIMATE statements (sas.com)   ... View more

Yes, I meant the "treatment" and not covariates in this case. ... View more

Thank you  Rob. It worked. Im so excited. thank you  so  very much😊 @SAS_Rob  ... View more

I'm sorry but I have one more question. If I am doing backward logistic regression I was told that I could not remove both of the variables with imputed values from the model. When I did this I kept getting a variance of zero. I guess my question is with backward approach logistic regression do I have to keep both of the variables with imputed values in the model, even if the P value is the least significant?   BA ... View more

Thanks for bailing me out @SAS_Rob    SteveDenham ... View more

Thanks for the reply, much appreciated.  I understood that the modified Poisson approach can be used (repeated statement), however GEE has the advantage of specifying the correlation structure as we see would fit the research question. Is this available with the modified Poisson approach? thank you   ... View more

Ah, yes!  Thanks, I was looking in the wrong place.   Regards, Brandon ... View more

Yes, you will get n-1 sets of coefficients, where n is the number of levels to the response variable.  This is because they are restricted in a sense that the probabilities must sum to 1.  There is a good explanation as to how to interpret a generalized logit model in general in the LOGISTIC documentation. SAS Help Center   ... View more

This section in the documentation will be helpful in determining which method you should use based on the missing data pattern and the type of variable. SAS Help Center: Imputation Methods   In your case you will likely want to decide whether to use either the DISCRIM or LOGISTIC method. ... View more

All these questions, including the rate ratio, are addressed in this note. It is illustrated using PROC GENMOD, but the same basic methods apply. ... View more