A survey of 200 adults showed that 15% played basketball for regular exercise. If a researcher desires to find the 95% confidence interval of the true proportion of adults who play basketball and be within 1% of the true population, how large a sample should be selected?