Use X and β as defined for the completely random design, equation 9.4. Define K for the null hypothesis H 0 : τ 1 = τ 2 . Define K for the null hypothesis H 0 : τ 3 = τ 4 . Define K for the composite...

Use X and β as defined for the completely random design, equation 9.4. Define K for the null hypothesis H₀: τ₁
= τ₂. Define K for the null hypothesis H₀: τ₃
= τ₄. Define K for the composite null hypothesis H₀: τ₁
= τ₂
and τ₃
= τ₄
and τ₁
+ τ₂
= τ₃
+ τ₄. Is each of these hypotheses testable? How does the sum of squares generated by the composite hypothesis relate to the analysis of variance?