P15-25. High-Low and Scatter Diagrams with Implications for Regression

Midnight Cookie Company produces and delivers gourmet cookies and ice cream until 1:30 a.m. from its three Seattle area locations. Presented is monthly cost and sales information for cookies at one of Midnight's locations.

Month

Sales (Dozens)

Total Costs

January

6,800

30,650

February

7,800

35,336

March

5,500

29,700

April

1,000

25,000

May

6,100

30,600

June

4,500

28,670

Required

a. Using the high-low method, develop a cost-estimating equation for total monthly costs.

b. 1. Plot the equation developed in requirement (a).

2. Using the same graph, develop a scatter diagram of all observations for the cookie shop. Select representative high and low values and draw a second cost-estimating equation.

c. Which is a better predictor of future costs? Why?

d. If you decided to develop a cost-estimating equation using least-squares regression analysis, should you include all the observations? Why or why not?

e. Mention two reasons that the least-squares regression is superior to the high-low and scatter diagram methods of cost estimation.