I can't say with 100% certainty what is going on without having access to your data, but it looks like you are getting the results of two different tests of the null hypothesis that the variance of the random effects is zero. The default for the covtest statement is a likelihood ratio test, and then you have also requested a Wald test. These two types of tests are not guaranteed to produce the same results. You can read more about them in the 'Statistical Inference for Covariance Parameters' section of the documentation (https://go.documentation.sas.com/doc/en/pgmsascdc/9.4_3.4/statug/statug_glimmix_details27.htm ). You'll note that there is subsection that specifically discusses likelihood ratio vs. Wald tests. In general, likelihood ratio tests are often preferred to Wald tests.  ... View more

I am able to run PROC HPGENSELECT on either SAS 9.4 or SAS Studio. For some example code that I can run successfully, see here: https://go.documentation.sas.com/doc/en/statcdc/14.3/stathpug/stathpug_code_hpgenex1.htm . ... View more

You can specify the random effects either way; a single RANDOM statement that contains both int and day, or with a separate RANDOM statement for each effect.  ... View more

This looks like you are on the right track. On quick review, I have two thoughts. First, consider whether you want to include a random intercept in the model. It's not necessarily required, just think about whether you expect random variation in the outcome variable at baseline, and whether you want to capture that variation in your model. Second, I can't tell exactly how your data is structured, but you might need to modify the subject= argument in the random statement. If it's the case that each pen has its own unique identifier, then what you have looks fine. But if the pen numbers are nested within block, i.e., there is a pen numbered 1 in block 1, block 2, etc., then consider subject = pen(block).  ... View more

This is closely related to the LSMEANS statement, but you can also have a look at the syntax for the 'slice' statement. It looks to me like something along the lines of   slice reg_use*CIG6C / sliceBy = CIG6C diff exp;  may get you going in the right direction.  ... View more

The question of whether/how the results of your analysis will generalize beyond your study is going to largely depend on how representative these movies are of the larger collections of movies in your population of interest. Which is to say - it mainly depends on how you selected the movies that you studied. It's not the case that, if you use random effects, you can generalize your results, and otherwise you can't.  ... View more

I'm seeing the same kind of behavior. I'm not getting an error, but I stopped the command after a few minutes of run-time. I'm not sure if it would eventually converge if I let it go indefinitely, but, regardless, I think the issue is that this is just a very time and memory intensive model to fit. If I use, for example, only the random statement or only the repeated statement, the model converges reasonably quickly. Of course, convergence does not mean that the model is appropriate for the data. All of which is to say, to model your data, I would recommend trying to simplify model fitting by accounting for correlation among observations by using either the random statement (i.e., mixed effects model) or the repeated statement, but not both.  ... View more

Can you post the code that is causing the issue?  ... View more

As @Rick_SAS alluded to, logistic regression is probably ok to use if all you have is 1-year survival (yes/no). However, if you actually have survival times for each person (e.g., person 1 survived 5 years, person 2 survived 6 months, person 3 was lost to follow-up at 3 years), you should generally not use logistic regression and instead prefer a model tailored to survival data. One possibility is proportional hazards regression, which can be implemented using PROC PHREG.    To your question about categorizing continuous variables, you should almost always avoid doing that. Rather, use a model that can accommodate continuous predictors.  ... View more

To the previous comment, I would add that, if you decide on modeling the outcome as multinomial, you could also consider proc GEE. More generally, I would suggest that if you decide on using random effects, have a specific reason for doing so. For instance, you might be interested in estimating subject-specific effects (rather than population-averaged effects), or you might be interested in analyzing the variance of the random effects, etc. If you are not specifically interested in using the random effects, you do not, in general, need to use random effects to account for correlation among the responses. You can instead use methods that allow you to directly specify the working correlation structure. In proc GEE and proc MIXED, you can do so by using the 'repeated' statement.   ... View more

It would be helpful to provide a little more context for this question, but my reading of this is that you have submitted a manuscript for publication and you are posting a portion of the reviewer comments. If that is correct, I think the reviewer is asking for information about how you arrived upon the number of participants that you enrolled in the study. They are assuming that, to come up with a target sample size, you conducted a power analysis *prior* to the study. The reviewer is asking for the parameters of your power calculation (i.e., the effect sizes assumed in each group, variance, etc.). If you didn't do a power calculation *prior* to the study, you won't be able to fix that, you will just need to state that as a limitation. It sounds like what you reported were calculations based on the study data *after* the study had been completed, and they are asking you to remove that information.    As to the comment about multiple comparisons, many of the SAS stat procedures have options to adjustment for multiple tests. A starting point might be the MULTTEST procedure.  ... View more

I might not be completely understanding your question. What I gather is you want to report, separately for sex, race, clinic, etc., age-adjusted odds ratios of disease. In that case, you need to include age in your model statement, e.g.,  model Disease (event = "Yes") = sex age;  If age is a categorical measure, you also need to include age in your class statement.    ... View more