A medical researcher wishes to determine the percentage of females who take vitamins. He wishes to be 95% confident that the estimate is within 2.5 percentage points of the true proportion. A recent study of 320 females showed that 66% took vitamins. How large should the sample size be?