Sorrentino Company, which has been in business for one year, manufactures specialty Italian pastas. The pasta products start in the mixing department, where durum flour, eggs, and water are mixed to form dough. The dough is kneaded, rolled flat, and cut into fettucine or lasagna noodles, then dried and packaged. Paul Gilchrist, controller for Sorrentino Company, is concerned because the company has yet to make a profit. Sales were slow in the first quarter but really picked up by the end of the year. Over the course of the year, 717,500 boxes were sold. Paul is interested in determining how many boxes must be sold to break even. He has begun to determine relevant fixed and variable costs and has accumulated the following per unit data:

Price $0.90
Direct materials 0.35
Direct labor 0.25

He has had more difficulty separating overhead into fixed and variable components. In examining overhead-related activities, Paul has noticed that machine hours appear to be closely correlated with units in that 100 boxes of pasta can be produced per machine hour. Setups are an important batch-level activity. Paul has accumulated the following information on overhead costs, number of setups, and machine hours for the past 12 months:

Overhead Number of Setups Machine Hours

January $5,700 18 595
February 4,500 6 560
March 4,890 12 575
April 5,500 15 615
May 6,200 20 650
June 5,000 10 610
July 5,532 16 630
August 5,409 12 625
September 5,300 11 650
October 5,000 12 550
November 5,350 14 593
December 5,470 14 615

Selling and administrative expenses, all fixed, amounted to $180,000 last year.

Required:
Separate overhead into fixed and variable components using ordinary least squares (regression) analysis. Run three regressions, using the following independent variables: (a) number of setups, (b) number of machine hours, and (c) a multiple regression using both number of setups and machine hours. Which regression equation is best? Why?