If a researcher wants to estimate the proportion of a population with $95\%$ confidence ($z^* = 1.96$) and a margin of error $E=0.02$, assuming $p=0.5$, what is the required sample size?