To obtain the average partial effect of train using the probit model, we obtain tted probabilities for each man for train = 1 and train = 0. Of course, on of these is a counterfactual, because the man was either in job training or not. Of course, we evaluate the other regressors at their actual outcomes. The APE is the average, over all observations, of the differences in the estimated probabilities. With the variables in part (ii) appearing in the probit, the estimated APE is about - .112. Interestingly, rounded to three decimal places, this is the same as the coefficient on train in the linear regression. In other words, the linear probability model and probit give virtually the same estimated APEs