When we test for equality of population variances, we consider Alfa/2 in our decision criteria as it is a two sided test. But in IAI Nov 2008 Q13 a) they've taken Alfa only in finding the critical points.
Since alpha = 0.2 we should be using the 10% critical values from the F tables. I can't see how they could not divide by two as the 20% F tables are not given in the Tables...