9.39 In studying the proportion of water samples containing harmful bacteria, how many samples should be selected before and after a chemical is added if we want to estimate the true difference between proportions to within 0.1 with a 95% confidence coefficient? (Assume the sample sizes are to be equal.)