You want to obtain a sample to estimate a population proportion. At this point in time, you have no reasonable estimate for the population proportion. You would like to be 90% confident that you estimate is within 1.5% of the true population proportion. How large of a sample size is required?

Do not round mid-calculation.