Pierce Company manufactures drill bits. The production of the drill bits occurs in lots of 1000 units. Due to the intense competition in the industry and the correspondingly low prices, Pierce has undertaken a study of the manufacturing costs of each of the products it manufactures. One part of this study concerns the overhead costs associated with producing the drill bits. Senior production personnel have determined that the number of lots produced, the direct labor hours used, and the number of production runs per month might help to explain the behavior of overhead costs. The file P11_62.xlsx contains the data on these variables for the past 36 months.

a. How well can you can predict overhead costs on the basis of these variables with a linear regression equation? Why might you be disappointed with the results?

b. A production supervisor believes that labor hours and the number of production run setups affect overhead because Pierce uses a lot of supplies when it is working on the machines and because the machine setup time for each run is charged to overhead. As he says, "When the rate of production increases, we use overtime until we can train the additional people that we require for the machines When the rate of production falls, we incur idle time until the surplus workers are transferred to other parts of the plant. So it would seem to me that there will be an additional overhead cost whenever the level of production changes. I would also say that because of the nature of this rescheduling process, the bigger the change in production, the greater the effect of the change in production on the increase in overhead." How might you use this information to find a better regression equation than in part a?