Want to find the sample size required for alpha=0.05 and beta =0.05 (power 95%) in testing the equality of two normal population variances, when actually one variance is 25.5 times larger than the other. The right tail of the F distribution is chosen for the rejection region.

I believe this implies

Ho: sigmaSquared1 = sigmaSquared2 <-> (sigmaSquared1/sigmaSquared2) =1

Ha: sigmaSquared1 > sigmaSquared2 <-> (sigmaSquared1/sigmaSquared2) >1 *Since we are told using the upper tail.

The true situation must be sigmaSquared1 =25.5*sigmaSquared2

1) Do I infer correctly that since we are using the upper tail of the F that the sign of Ha is > ?

Now, how to get the sample size I am at a loss.

Any hints?