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, so we assume p=.5. You would like to be 99% confident that you estimate is within 0.2% of the true population proportion. How large of a sample size is required?