We want to estimate the actual difference between the proportions of non-smokers and smokers who said "yes" to federal tax increase. Notation: 1=non-smokers and 2=smokers. Based on this data, what is the upper bound for a 95% confidence interval for the difference in the population proportions