2. A medical sociologist takes a random sample of lower income residents in Texas to estimate the percentage of income (PI) spent on medical insurance. The distribu- tion of PI in Texas is is highly right skewed. She wants a 95% confidence interval on the median value of PI. You apply a Box-Cox transformation and obtain X = g(PI), where g(t) = 1/vt. A p-value = .456 is obtained from the Shapiro-Wilk statistic for the transformation data. Which of the following methods would you recommend for constructing the confidence interval? A. A studentized Bootstrap procedure B. Use the inverse of the normal- based tolerance interval: ([X + +(.025) Sx/ Vn] , [X -t(.025) Sx/ Vn] C. Use a nonparametric confidence interval D. Estimate the population pdf using a kernel density estimator and then obtain MLEs for the parame- ters in the kernel density estimator E. None of the above procedures would work very well ANS