ABSTRACT
Hence the likelihood ratio test rejects H0 whenever r 2=C, the constant C depends on
the level of significance a of the test. The distribution of R2 under H0 is the same as
that of R2 in the multivariate normal case (6.86) with If q in (8.230) is convex this test is uniformly most powerful invariant for testing H0 against H1:
?2>0. The proof is similar to that of Theorem 8.3.4. We refer to Giri (1988) and Kariya and Sinha (1989) for details and other relevent results.
