How SAS studio calculates BIC in proc mixed?

[johamunoza0]

Dear All,

I am trying to understand the way that BIC is calculated for a mixed model using PROC MIXED, I arrived to the  AIC reported by SAS under the ML and REML version.

However when i tried manually to calculate the BIC it seems that it is penalized with a function of log(n^2) instead of log(n), being n the number of subjects.

Could somebody iluminate me how  SAS calculates BIC under ML and REML? 

I am using the currently studio version,

Thanks a lot for your help 

[Reeza]

http://documentation.sas.com/?docsetId=statug&docsetVersion=14.2&docsetTarget=statug_mixed_syntax01....

[johamunoza0 · Accepted solution]

Thanks so much for your reply, i had not considered  the 2 in the AIC complexity penalty.  Best wishes, Johanna