29. When sample sizes are equal (Ji
J), the parameters

1, 2, . . . , I of the alternative parameterization are restricted by
i
0. For unequal sample sizes, the most natural restriction is
Ji

i
0. Use this to show that E(MSTr)
2
Ji

2 i

What is E(MSTr) when H0 is true? [This expectation is correct if
Ji

i
0 is replaced by the restriction
i
0

(or any other single linear restriction on the is used to reduce the model to I independent parameters), but

Ji

i
0 simplifies the algebra and yields natural estimates for the model parameters (in particular, ˆi
x i
 x

).]