Chapter 9 Multiple Regression 387
a company understate the calorie
content?
f. Save your changes to the workbook
and write a report summarizing your
observations.
- The Fritos workbook is a slight modi-
fi cation of the Wheat workbook with
data added about Fritos corn chips. It is
included because it has a substantial fat
content, in contrast to the other foods in
the data set. Because none of the foods
there have much fat, it is impossible to
see from the Wheat workbook how much
fat contributes to the calories in the foods.
a. Open the Fritos workbook from the
Chapter09 folder and save it as Fritos
Multiple Regression.
b. Repeat the regression of the previous
exercise and see whether the co-
effi cient for fat is now estimated more
accurately. Use both the known value
of 9 for comparison and the standard
error of the regression that is printed
in the output.
c. Save your changes to the workbook
and write a report summarizing your
observations.
- The Baseball workbook contains team
statistics for each of the major league
teams from the 2001–2007 baseball sea-
sons. You’ve been asked to derive an
equation that predicts the number of
runs per game on the basis of the num-
ber of singles, doubles, triples, home
runs, bases on balls, and strike outs.
a. Open the Baseball workbook from
the Chapter09 folder and save it as
Baseball Multiple Regression.
b. Create six new columns in the Base-
ball Stats worksheet and calculate the
average number of singles, doubles,
triples, home runs, bases on balls, and
strikeouts per game for each of the
teams in the data set.
c. Regress Runs per Game on the six
variables you created to derive an
equation for the average number of
runs per game on the basis of the
average number of singles, doubles,
triples, home runs, bases on balls,
and strike outs. Are all of the vari-
ables in your equation signifi cant?
Remove any insignifi cant variables
from your model and rerun the regres-
sion. Compare your results with the
results obtained by Rosner and Woods
(1988), as quoted in the beginning of
this chapter. Can the differences be
explained in terms of the standard
errors of the coeffi cients?
d. Do the Rosner-Woods coeffi cients
make more sense in terms of which
should be largest and which should
be smallest?
e. Save your changes to the workbook
and write a report summarizing your
results.
- The Toyota workbook contains price,
age, and mileage data for used car sales
of Toyota Corollas from 2009. You’ve
been asked to analyze the data to model
the effect of age and mileage on the used
car price.
a. Open the Toyota workbook from the
Chapter09 data folder and save it as
Toyota Multiple Regression.
b. Regress price on age and miles. What
impact do age and miles have on the
sale price of the car? Are both vari-
ables signifi cant in your regression
equation?
c. Create plots of Residuals versus
Miles, Residuals versus Age, and
Residuals versus Predicted Price.
Do you notice any pattern in
the graphs that would indicate a
problem with the constant vari-
ance assumption or the linearity
assumption?
