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

bottom-
up approach
(starting with a smaller model and expanding it as one goes along) or by
the somewhat pejorative term, 
data mining
(other names are 
regression fishing, data
grubbing, data snooping,
and 
number crunching
)
.
The primary objective of data mining is to develop the “best” model after several diag-
nostic tests so that the model finally chosen is a “good” model in the sense that all the
estimated coefficients have the “right” signs, they are statistically significant on the basis of
the 
t
and 
F
tests, the 
R
2
value is reasonably high, and the Durbin–Watson 
d
has acceptable
value (around 2), etc. The purists in the profession look down on the practice of data
mining. In the words of William Pool, “. . . making an empirical regularity the foundation,
rather than an implication of economic theory, is always dangerous.”
15
One reason for
“condemning” data mining is as follows.
Nominal versus True Level of Significance in the Presence of Data Mining
A danger of data mining that the unwary researcher faces is that the conventional levels of
significance (
α
) such as 1, 5, or 10 percent are 
not the true levels of significance
. Lovell has
suggested that if there are 
c
candidate regressors out of which 
k
are finally selected (
k
≤
c
)
on the basis of data mining, then the true level of significance (
α
*
) is related to the nominal
level of significance (
α
) as follows:
16
α
∗
=
1
−
(1
−
α
)
c
/
k
(13.4.2)
15
William Pool, “Is Inflation Too Low?” the 
Cato Journal,
vol. 18, no. 3, Winter 1999, p. 456.
16
M. Lovell, “Data Mining,” 
Review of Economics and Statistics,
vol. 65, 1983, pp. 1–12.
guj75772_ch13.qxd 16/08/2008 03:24 PM Page 475

---

476