11.6 A Lack-of-Fit Analysis
A lack-of-fit analysis can provide information on the adequacy of a model that does not include all possible terms. The basic principle is to compare the fits of “complete” and “reduced” models as described in Chapter 2. The sum of squares for the complete model can be obtained from a one-way analysis of variance that computes the sums of squares among all unique treatments. The error sum of squares for the full model is subtracted from the error sum of squares for the incompletely specified model to obtain the sum of squares for all terms not specified.
One of the most common applications of lack-of-fit analysis is testing the adequacy of a regression model. In this procedure, you want to determine if a fitted model ...
Get SAS for Linear Models, Fourth Edition now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
