Page 112 - Intermediate Statistics for Dummies

10_045206 ch05.qxd 2/1/07 9:49 AM Page 91
Chapter 5: When Two Variables Are Better than One: Multiple Regression
Advertising Dollars and Sales of Plasma TVs
Location
20
1
0
9.73
11.19
2
0
20
5
3
5
8.75
5
5
4
6.25
9.10
5
10
10
10
9.71
6
10
7
15
9.31
15
15
11.77
8
15
9 Table 5-1 Sales ($ mil) TV Ads ($1,000) Newspaper Ads ($1,000) 91
5
20
8.82
10 9.82 20 5
11 16.28 25 25
12 15.77 25 25
13 10.44 30 0
14 9.14 30 0
15 13.29 35 5
16 13.30 35 5
17 14.05 40 10
18 14.36 40 10
19 15.21 45 15
20 17.41 45 15
21 18.66 50 20
22 17.17 50 20
The question is, can the amount of money spent on these two forms of adver-
tising do a good job of estimating sales (in other words, are the ads worth the
money)? And if so, do you need to include spending for both types of ads to
estimate sales, or is one of them enough? Looking at the numbers in Table 5-1,
you can see that higher sales may be related at least to higher amounts spent
on TV advertising; the situation with newspaper advertising may not be so
clear. So will the final multiple regression model contain both x variables or
only one? In the following sections, you can find out.