Suppose we wish to design a study to determine the proportion of patients who received pain therapy following a particular surgical procedure in a local hospital. How many subjects would be required in the study to ensure that the estimate was within 4 percentage points of the true proportion with 95% confidence?