Using G*Power assume that you have a sample size beyond what you can obtain. Use the compromise function to compute an alpha and beta sample half the size of your original size. Indicate the resulting alpha and beta. Is the study worth doing with a smaller sample?