In independent samples t-tests, power is a function not only of N, but also of how close to equal the ns are in each group. Compute a table showing the power for effect sizes of 0.4, 0.6, and 0.8 when the ns for the two groups are: 10 and 50, 20 and 40, and 30 and 30.

Which of the three effect sizes is most affected by the N not being equal? Why?