Ritter Manufacturing Company has kept track of machine hours and overhead costs at its main manufacturing plant for the past 52 weeks. The data appear in the file P08_45.xlsx. Ritter has studied these data to understand the relationship between machine hours and overhead costs. Although the relationship is far from perfect, Ritter believes a fairly accurate prediction of overhead costs can be obtained from machine hours through the equation

Estimated Overhead 5 746.5078 1 3.3175* Machine Hours

By substituting any observed value of Machine Hours into this equation, Ritter obtains an estimated value of Overhead, which is always somewhat different from the true value of Overhead. The difference is called the prediction error.
a. Calculate a 95% confidence interval for the mean prediction error. Do the same for the absolute prediction error. (For example, the prediction error in week 1, actual overhead minus predicted overhead, is 994.5303. The absolute prediction error is the absolute value, 94.5303.)
b. A close examination of the data suggests that week 45 is a possible outlier. Illustrate this by creating a box plot of the prediction errors. In what sense is week 45 an outlier? See whether week 45 has much effect on the confidence intervals from part a by recalculating these confidence intervals, this time with week 45 deleted. Discuss your findings briefly.