Using data from Major League Baseball players,
you can estimate a model using the natural log of the player’s
contract value as the dependent variable and several player
characteristics as independent variables, including 3-year
averages for slugging percentages and at-bats, player age, and
player tenure on his current team. You can use STATA to
perform WLS by using a standard FGLS procedure. Figure
11-9 shows the resulting output. As you can see, the WLS
estimates are similar to the OLS estimates, but they’re not
identical. All the coefficients have the same sign before and
after the heteroskedasticity correction. In this example, all the
coefficients that were statistically significant with OLS
remain significant with WLS. However, the effect of tenure is
marginally significant with WLS but wasn’t significant with
OLS.