9781118041581

Regression Analysis 139

MULTIPLE REGRESSION Because price is not the only factor that affects sales,

it is natural to add other explanatory variables to the right-hand side of the

regression equation. Suppose the airline has gathered data on its competi-

tor’s average price and regional income over the same four-year period. In

management’s view, these factors may strongly affect demand. Table 4.4 lists

the complete data set. Management would like to use these data to estimate

a multiple-regressionequation of the form

In this equation, quantity depends on own price (P), competitor’s price (P),

and income (Y). Now the OLS regression method computes four coefficients:

the constant term and a coefficient for each of the three explanatory variables.

As before, the objective is to find coefficients that will minimize SSE. The OLS

equation is

Q28.842.12P1.03P3.09Y. [4.3]

QabPcPdY.

TABLE 4.4

Airline Sales, Prices,

and Income

Year and Average Number Average Average Average

Quarter of Coach Seats Price Competitor Price Income

Y1 Q1 64.8 250 250 104.0

Q2 33.6 265 250 101.5

Q3 37.8 265 240 103.0

Q4 83.3 240 240 105.0

Y2 Q1 111.7 230 240 100.0

Q2 137.5 225 260 96.5

Q3 109.5 225 250 93.3

Q4 96.8 220 240 95.0

Y3 Q1 59.5 230 240 97.0

Q2 83.2 235 250 99.0

Q3 90.5 245 250 102.5

Q4 105.5 240 240 105.0

Y4 Q1 75.7 250 220 108.5

Q2 91.6 240 230 108.5

Q3 112.7 240 250 108.0

Q4 102.2 235 240 109.0

c04EstimatingandForecastingDemand.qxd 9/5/11 5:49 PM Page 139