Regression Analysis and Cost Estimation:
The CEO of Carson Company has asked you to develop a cost equation to predict monthly overhead costs in the production department. You have collected actual overhead costs for the last 12 months, together with data for three possible cost drivers, number of Indirect Workers, number of Machine Hours worked in the department and the Number of Jobs worked on in each of the last 12 months:
|
Overhead Costs
|
Indirect Workers
|
Machine Hours
|
Number of Jobs
|
|
$2,200
|
30
|
350
|
1,000
|
|
4,000
|
35
|
500
|
1,200
|
|
3,300
|
50
|
250
|
900
|
|
4,400
|
52
|
450
|
1,000
|
|
4,200
|
55
|
380
|
1,500
|
|
5,400
|
58
|
490
|
1,100
|
|
5,600
|
90
|
510
|
1,900
|
|
4,300
|
70
|
380
|
1,400
|
|
5,300
|
83
|
350
|
1,600
|
|
7,500
|
74
|
490
|
1,650
|
|
8,000
|
100
|
560
|
1,850
|
|
10,000
|
105
|
770
|
1,250
|
(a) The CEO suggests that he has heard that the high-low method of estimating costs works fairly well and should be inexpensive to use. Write a response to this suggestion for the CEO indicating the advantages and disadvantages of this method. Include the calculation of a cost equation for this data using Machine Hours as the cost driver.
(b) Using Excel develop three scatter diagrams showing overhead costs against each of the three proposed independent variables. Comment on whether these scatter diagrams appear to indicate that linearity is a reasonable assumption for each.
(c) Using the regression module of Excel's Add-in Data Analysis, perform 3 simple regressions by regressing overhead costs against each of the proposed independent variables. Show the output for each regression and evaluate each of the regression results, indicating which of the three is best and why.
Provide the cost equations for those regression results which are satisfactory and from them calculate the predicted overhead in a month where there were 100 Indirect Workers and 500 Machine Hours and 1,000 Jobs worked.
(d) Selecting the two best regressions from part (c) conduct a multiple regression of overhead against these two independent variables. Evaluate the regression results.
Draw conclusions about the best of the four regression results to use.