You are the management accountant of the Sainsbury College, and are required to use regression analysis to estimate the cost function of asset repairs and maintenance for the College. You identify two possible cost drivers for the asset repairs and maintenance cost: number of maintenance hours and number of students.

Based on the past ten semesters' data, you first regress the asset repairs and maintenance cost on number of maintenance hours and the adjusted R-square is 70%, and then you regress the asset repairs and maintenance cost on number of students and the adjusted R-square is 31%. You also know that multiple cost drivers can be added into one regression model, and regress the cost of asset repairs and maintenance on both cost drivers. The regression output is as follows:

Coefficients t Stat P-value Intercept 10,000.00 1.45 0.25 Number of maintenance hours 38.00 3.19 0.005 Adjusted R-Square: 70% Coefficients t Stat P-value Intercept 55,000.00 2.44 0.08 Number of students 21.00 2.85 0.08 Adjusted R-Square: 31% Required: I 1. If only one cost driver is used, which cost driver, number of maintenance hours or number of students, is a better predictor and why? 2. Create a cost function for each of the cost drivers. Using the appropriate cost function based on your answer above, predict the cost of asset repairs and maintenance if the number of maintenance hours is predicted to be 200 and number of students is predicted to be 350 in the next semester. 3. "Regression analysis increases cost function accuracy by using all points." Can managers always rely on the results of regression analysis and should it always be given preference over other cost estimation technique? " Focus