if you want to find a confidence interval for the mean weight loss, is it necessary to know that the population distribution of weight loss is normally distributed? explain why or why not.