Parameter estimates resulting from analysis of variance
models where the effects are all classification variables
are different from those observed in a regression model.
They represent a non-unique solution to the normal equations,
and thus the individual elements in the table are not as
easily interpretable as they are in multiple regression.
For a complete discussion of parameter estimates involving
classification variables, refer to the chapter
"Details of the Linear Model: Understanding GLM Concepts,"
in SAS System for Linear Models, Third Edition.
Figure 15.8: Fit Window - Parameter Estimates
Copyright © 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.