Dr. Paul Oswiecmiski wants to estimate the mean serum HDL cholesterol of all 20- to 29-year-old males. How many subjects are needed to estimate the mean serum HDL cholesterol of all 20- to 29-year-old males within 1.5 points with 90% confidence, if s=12.5 based on earlier studies? Suppose that Dr. Oswiecmiski would prefer 95% confidence. How does the increase in confidence affect the sample size required?