1. The two-way ANOVA, non-orthogonal case, has been a vexing problem for ANOVA researchers for many years.  Please answer the following questions concerning the two-way non-orthogonal ANOVA.

a. Explain the meaning of the word non-orthogonal.

b. What condition(s) must exist for non-orthogonality to occur?  Be specific. 

c. Why is the non-orthogonal ANOVA more difficult than the orthogonal ANOVA?

d. What is a Type I effect for the main effects of A and B in terms of model comparisons?

e. What is a Type II effect for the main effects of A and B in terms of model comparisons?

f. What is a Type III effect for the mains effects of A and B in terms of model comparisons?

g. What assumption is made for a Type II effect or Type II means to be legitimate?

h. What assumption is made for a Type I effect or a Type I mean to be legitimate?

i. What is the equivalent problem in regression?