Unformatted text preview: Mr. Ladouceur should not use the specification derived from the linear regression. The chart of the residuals shows that the error terms e are not randomly scattered around the mean (the horizontal line). After the initial increase from January to February, there is a steady downward drift to the values. The residuals also appear to have a wave form. This is further evidence that a reliable cost driver has not yet been specified. Residuals Y - y = e Residual Plot
-40000 2,000 4,000 6,000 8,000 Labour Hours 10‐29 1. 2. 3. Evaluating multiple regression models, not‐for‐profit It is economically plausible that the correct form of the model of overhead costs includes both number of academic programs and number of enrolled students as cost drivers. The findings in Collaborative Learning Case 10‐28 indicate that each of the independent variables affects overhead costs. (Each regression has a significant r2 and t‐value on the independent variable.) Hanks could choose to divide overhead costs into two cost pools, (i)...
View Full Document
- Winter '12
- overhead costs, Mr. Ladouceur