Common Statistical Methods for Clinical Research with SAS Examples, Third Edition, 3rd Edition

Appendix D. SS Types I, II, III, and IV Methods for an Unbalanced Two-Way Layout

• D.1 SS Types Computed by SAS 457

• D.2 How to Determine the Hypotheses Tested 459

• D.3 Empty Cells 464

• D.4 More Than Two Treatment Groups 465

• D.5 Summary 468

D.1 SS Types Computed by SAS

This section uses the randomized block design to illustrate the differences among the various ANOVA types of sums of squares (SS) computed by SAS: Type I, Type II, Type III, and Type IV.

Consider the unbalanced two-way layout that has two treatment groups (A and B) and 3 study centers (1, 2, and 3). Let μ_ij and n_ij represent the mean and sample size, respectively, for Treatment i, Center j. For this example, 8 patients receive Treatment A and 6 patients receive Treatment B. The patients are ...