The editor ofNew Yorkstudies information on 101 various undergraduate business programs in U.S. and wants to explore the relationship between the median starting salary (in thousand dollars) of the graduates and the following set of potential explanatory variables: annual cost (in thousand dollars) of the program, full-time enrollment, faculty-student ratio, and average ACT score. (For example, median starting salary of $60,000 median.starting.salary=60)

She fits a multiple regression model and uses backwards elimination with =0.10 to achieve a parsimonious model. Below are the Excel outputs from her first and last models.

(a) What was her initial proposed model? What explanatory variables did she remove? Why did she do so?

(b) What is the final fitted model?

(c) Calculate and interpretR2of the final fitted model.

(d) Predict the median starting salary of a business program with annual cost of $37,000 (annual.cost=37) and average ACT score of 30 (average.ACT.score=30).

(e) One of the explanatory variables in the final fitted model has a p-value of 0.088. What hypothesis test does this p-value reference? What decision do you make about this test? What does it mean to retain the null? What does it mean to reject the null?

(f)What is the test statistic value that corresponds to this p-value of 0.088?

(g) The editor claims that she conducted residual analysis and that the model assumptions appear reasonable. What model assumptions should the editor have checked at the end of model building?

\f\f