166 BUSINESS APPLICATIONS OF MULTIPLE REGRESSION
Tolerance
1 – r
2
for this reduced multiple regression run. It is the case that the smaller
the tolerance, the greater the multicollinearity regarding the indepen-
dent variable being used as the dependent variable. The smallest possible
value for tolerance is zero, and a good rule of thumb is that anything
below 0.20 indicates a problem with multicollinearity. Two notes are in
order. First, with k independent variables, there will be k measures of
tolerance, as each independent variable is used, in turn, as the dependent
variable. Second, tolerance requires at least three independent variables
because with just two independent variables, the correlation coefficient is