Hi all, I am trying to run a multiple linear regression in either Proc MIXED and am experiencing problems. My DV, Y, is fairly right-skewed, and can be seen here: However, I'm not sure that is necessarily an issue with Proc MIXED, and moreover I don't believe it's the cause of my errors below (please let me know if this assumption is incorrect). I am trying to run a regression in proc mixed using the following code: proc mixed data = Data method = ML; class X8 (ref = '1'); model Y = X1 X2 X3 X4 X5 X6 X7 X8 ; run; Since this is not a longitudinal analysis, I am neither interested in random effects nor repeated measures. I am simply trying to use Proc Mixed to use maximum likelihood estimation to avoid listwise deletion found in proc reg etc. However, I am receiving this feedback upon running the model: NOTE: 756 observations are not included because of missing values. NOTE: An infinite likelihood is assumed in iteration 0 because of a nonpositive residual variance estimate. NOTE: PROCEDURE MIXED used (Total process time): real time 0.32 seconds cpu time 0.04 seconds I am not sure how to proceed with troubleshooting this or where to begin exploring the data to determine the issue. I read over a correlation matrix and there aren't any exceptionally correlated variables (the highest magnitude I observed was approximately .45). There is significant missing in some of the variables (some of them are extremely missing), but I'm not aware that this would cause the issue, given my understanding that Proc MIXED should use ML estimation to provide valid estimates. In case it helps, here is a snapshot of the missingness of the variables: Variable N N Miss Y 323 438 X 757 4 X2 321 440 X3 321 440 X4 761 0 X5 384 377 X6 11 750 X7 752 9 X8 761 0 Can anybody please advise on how to proceed diagnosing the issue? I am happy to provide further information if it's helpful (the distribution of the independent variables, etc.) Anything would be greatly appreciated. Thanks, Luke
... View more