You want to obtain a sample to estimate a population proportion. At this point in time, you have no reasonable preliminary estimation for the population proportion. You would like to be 99% confident that you estimate is within 3.5% of the true population proportion. How large of a sample size is required?
