Suppose that we have a oneway ANOVA with equal numbers of observations on each treatment, that is, ni = n, i = 1,..., k. In this case the F test can be considered an average t test.
(a) Show that a t test of H0: θi = θi€² versus H1: θi ‰  θi€² can be based on the statistic

(b) Show that

where F is the usual ANOVA F statistic.

Transcribed Image Text:

(Y - Y,)2