Yesterday (old test question) there was a referendum in Italy. The measure would only count if 50% of the voters turned out to vote. The measure passed. But unfortunately for its backers, the turnout was only 49.6%. This percentage is not a sample, because all the ballots were counted. If instead of counting all the ballots, you had to determine the percentage of Yes votes with a sample, how large a sample would you have to take to get the result within 0.2% at 95% confidence? (0.3 points)