A pollster would like to estimate the fraction p of people in a popu- lation who intend to vote for the particular candidate. How large must a random sample be in order to be at least 95% certain that the fraction ^p of the positive answers in the sample is within 0:02 of the true p?