The GLIMMIX Procedure

Between-Within Degrees of Freedom Approximation

The DDFM=BETWITHIN option divides the residual degrees of freedom into between-subject and within-subject portions. PROC GLIMMIX then determines whether a fixed effect changes within any subject. If so, it assigns within-subject degrees of freedom to the effect; otherwise, it assigns the between-subject degrees of freedom to the effect (Schluchter and Elashoff 1990). If the GLIMMIX procedure does not process the data by subjects, the DDFM=BETWITHIN option has no effect. See the section Processing by Subjects for more information.

If multiple within-subject effects contain classification variables, the within-subject degrees of freedom are partitioned into components that correspond to the subject-by-effect interactions.

One exception to the preceding method is the case where you model only R-side covariation with an unstructured covariance matrix (TYPE=UN). In this case, all fixed effects are assigned the between-subject degrees of freedom to provide for better small-sample approximations to the relevant sampling distributions. The DDFM=BETWITHIN method is the default for models with only R-side random effects and a SUBJECT= option.