The McGraw-Hill Series Economics essentials of economics brue, McConnell, and Flynn Essentials of Economics

TABLE 11.5
Sales and
Employment 
for Companies
Performing
Industrial R&D 
in the United States,
by Industry, 2005
(values are in
millions of dollars)
Source: National Science
Foundation, Division of
Science Resources Statistics,
Survey of Industrial Research
and Development: 2005 and
the U.S. Census Bureau
Annual Survey of
Manufacturers, 2005.
To see if the regression (11.7.3) suffers from heteroscedasticity, we obtained the resid-
uals, 
ˆ
u
i
, and the squared residuals, 
ˆ
u
2
i
, from the model and plotted them against sales, as
shown in Figure 11.13. It seems from this figure that there is a systematic pattern between
the residuals and squared residuals and sales, perhaps suggesting that there is het-
eroscedasticity. To test this formally, we used the Park, Glejser, and White tests, which gave
the following results:
Park Test
ˆ
u
2
i
= −
72,493,719 
+
916.1 Sales
i
se
=
(54,940,238)
(303.9)
(11.7.4)
t
=
(
−
1.32) (3.01)
r
2
=
0
.
431
The Park test suggests that there is a statistically significant positive relationship between
squared residuals and sales.
(
Continued
)
guj75772_ch11.qxd 14/08/2008 10:01 AM Page 397

---

398
Part Two
Relaxing the Assumptions of the Classical Model
0
–20,000
100,000
200,000
Sales
300,000
400,000
–10,000
10,000
0
20,000
Residuals
30,000
0
0
100,000
200,000
300,000
Sales
400,000
100,000,000
200,000,000
Residuals squared
300,000,000
400,000,000
500,000,000