It is believed 30% of all adults are overweight. How larger sample is necessary to estimate the true proportion of adults who are overweight with a 95% confidence interval so that the margin of error of the estimate is no greater than 2 percentage points?