*
*  Example 18.2
*  J. Wooldridge, Econometrics of Cross Section and Panel Data
*
*  Note: The sample data files include all required data transformations.
*        However, in many cases we show how these data transformations would
*        be created in RATS from the basic data, as this is something you will
*        need to know how to do in practice.
*
*
open data jtrain2.raw
data(format=free,org=columns) 1 445 train age educ black hisp married nodegree mosinex $
  re74 re75 re78 unem74 unem75 unem78 lre74 lre75 lre78 agesq mostrn
*
* Do first stage probit of training on the explanatory variables. Get the
* predicted probabilities.
*
ddv(dist=probit) train
# constant re74 re75 age agesq nodegree married black hisp
prj(cdf=propensity)
*
* Regress real earnings on intercept, train and the propensity variable.
*
linreg re78
# constant train propensity
*
* For comparison, run a linreg with the dummy and all the explanatory variables.
*
linreg re78
# constant train re74 re75 age agesq nodegree married black hisp
*
* And just the dummies.
*
linreg re78
# constant train
*
* Generate the interaction term for regression 18.24. The regression won't match
* exactly, because this doesn't round the mean to .416.
*
stats(noprint) propensity
set wp = train*(propensity-%mean)
*
linreg re78
# constant train propensity wp