Diamond State is trying to develop a cost formula for its major manufacturing overhead activities. The company's manufacturing process is highly automated and power costs are a significant manufacturing cost. Cost analysts have decided that power costs are mixed. The costs must be separated into their fixed and variable components so that the cost behavior of the power usage activity can be better understood.

Period	Machine Hours	Power Cost
Jan	24,000	$32,500
Feb	30,000	47,500
Mar	36,000	53,125
Apr	26,400	46,250
May	25,200	42,500
June	21,600	36,250
July	28,800	45,000
Aug	33,600	50,000
Sept	31,200	40,000

(1) Using the high and low points (high-low method) to develop and show the formula to predict yearly power cost. Formula should be in the form: Y = A + b(x), when Y is the total predicted power cost for the nine months, A is the total fixed cost, b is the variable power cost per machine hour, and x is the number of machine hours.

(2) Prepare a scattergraph.

(3) If Diamond State estimates machine hours of 29,500 for October, compute the power rate per machine hour using the high low method formula and least squares formula.

(4) At any level of activity within relevant range, if activity goes up by 2000 machine hours, how much would we expected the estimated power costs to increase using the regression method?

(5) Change the power costs for August to $60,000. What is your new formula for predicting power costs using least squares

(6) What effect would this change in August have on the predictive formula using the high low method?

(7) Which of the two methods is preferable?