Question: The belting department of a Cleveland based Belting co. functions primarily as a job shop, fabricating industrial belting for the users of belting in Northern Ohio. Many different types of belting are purchased in bulk (slab) quantities and are prepared to meet the specifications of each customer. At present, pricing of these custom jobs are based on material cost plus an estimated labor cost which is computed from estimated manufacturing time. Estimating the manufacturing time is a constant problem. The company would like to see if a regression model can be used to estimate the time more accurately. The Director of Manufacturing thinks that the manufacturing time depends on the belt size (length) and whether a splice to join the ends of the belt together is needed or not. Table below provides information about recent customer orders in which TYPE is an indicator variable with value 1 if a splice is needed and 0 otherwise. (You can download this file BELT1.XLS from Blackboard).

MAN_TIME LENGTH TYPE

(Hours) (Yards) (splice/no splice)

CASE 1 60 268 1

CASE 2 20 30 0

CASE 3 4 12 0

CASE 4 32 150 0

CASE 5 20 150 0

CASE 6 11 10 0

CASE 7 25 70 0

CASE 8 22 4 1

CASE 9 12 50 0

CASE 10 8 25 0

CASE 11 10 8 1

CASE 12 15 4 1

CASE 13 15 12 0

CASE 14 22 50 0

CASE 15 23 50 0

CASE 16 27 8 1

CASE 17 6 4 1

CASE 18 35 200 0

CASE 19 5 25 0

CASE 20 9 20 0

CASE 21 25 2 1

(a) Use EXCEL to develop a regression model. Interpret the meaning of all regression coefficients.

(b) What is R² and Standard Error of this model? Interpret these values in the output.

(c) Do both variables have significant incremental value? Discuss why or why not.

(d) The company has just received an order to supply a conveyor belt of length 250 yards with no joint (splice) needed. Based on this regression model, what is your estimate for the manufacturing time? How good you think is this estimate?