The use of Twitter for marketing political candidates has increased dramatically over the past several years. The Journal of Marketing Research (December 2019) published a study comparing tweeting patterns during and after a primary or presidential debate. The researchers analyzed over 9.5 million tweets posted during and shortly after several key debates leading up to the 2016 presidential election. One variable measured for each tweet was whether or not the tweet contained any visual content (e.g., a video, photo, or animated GIF). The researchers compared the percentage of tweets with visual content during the debates (estimated at 6%) with the percentage after the debates (estimated at 8%). Rather than analyze all 9.5 million tweets, you want to use a smaller sample to estimate the true difference in percentages. How many tweets are required to sample in order to estimate the difference to within 1% using a 95% confidence interval? Assume equal sample sizes and use the percentages estimated by the study.