For the following requirements, round the fixed cost to the nearest dollar and round the variable rates to the nearest cent.
1. Prepare a scattergraph by plotting power costs against machine hours. Does the scattergraph show a linear relationship between machine hours and power cost?
2. Using the high and low points, compute a power cost formula.
3. Use the method of least squares to compute a power cost formula. Evaluate the coefficient of determination.
4. Rerun the regression, and drop the point (20,000, $26,000) as an outlier. Compare the results from this regression to those for the regression in Requirement 3. Which is better?
The overall relationship looks reasonably linear—although the data point for the first quarter may be an outlier.
2. Using the high-low method:
Variable power cost = [pic] = $1.13 (rounded)
Fixed power cost = $42,500 – ($1.13 × 30,000) = $8,600
Total power cost = $8,600 + ($1.13 × Number of machine hours)
3. Output of regression program:
|SUMMARY OUTPUT | | | | | | | |
| | | | | | | | |
|Multiple R |0.89336