The illegible subscripts on the sigmas (between the first two horizontal rules) are epsilons (error).
The estimated variances (sigma hats on the right, below the first horizontal rule) are obtained from the expected mean squares on the left. You can practice obtaining them yourself.
Omega-hat squared is expressed in terms of the estimated variances, and also in terms of the F ratios from the omnibus ANOVA. You can show that the two forms are equivalent using the expressions for the estimated variances that appear between the first two horizontal rules.
Phi is expressed in terms of the estimated variances and also in terms of the omega-hat squareds. Again, you can show that the two forms are equivalent using the expressions that appear on this page.
Notice that everything here is analogous to the corresponding quantities for the one-way ANOVA.
Click your browser's back button to return to the Factorial ANOVA page.