A group of marketing students at a large university wants to determine the proportion of first-year students who use certain types of social media. The students want their estimate to be within 0.03 of the true proportion with a 90% level of confidence. How large of a sample is required if the population proportion in not known?