HERE IS THE STATA CODE I USED TO RUN THE NECESSARY REGRESSIONS: log using a:collegereg, replace t infile college ability using a:collegedata.txt regress college ability probit college ability dprobit college ability logit college ability log close clear ********************** LINEAR PROBABILITY MODEL ESTIMATES ********************* . regress college ability Source | SS df MS Number of obs = 1217 -------------+------------------------------ F( 1, 1215) = 420.85 Model | 76.5795109 1 76.5795109 Prob > F = 0.0000 Residual | 221.088525 1215 .181965864 R-squared = 0.2573 -------------+------------------------------ Adj R-squared = 0.2567 Total | 297.668036 1216 .244792793 Root MSE = .42657 ------------------------------------------------------------------------------ college | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- ability | .2509594 .0122333 20.51 0.000 .2269587 .27496 _cons | .4264668 .0122278 34.88 0.000 .4024767 .4504568 ------------------------------------------------------------------------------ ******************************8 PROBIT MODEL ESTIMATES ******************************* . probit college ability Iteration 0: log likelihood = -830.34831 Iteration 1: log likelihood = -652.46872 Iteration 2: log likelihood = -640.95584 Iteration 3: log likelihood = -640.78572 Iteration 4: log likelihood = -640.78568 Probit estimates Number of obs = 1217 LR chi2(1) = 379.13 Prob > chi2 = 0.0000 Log likelihood = -640.78568 Pseudo R2 = 0.2283 ------------------------------------------------------------------------------ college | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- ability | .8881127 .0531221 16.72 0.000 .7839954 .99223 _cons | -.2893961 .0418249 -6.92 0.000 -.3713714 -.2074208 ------------------------------------------------------------------------------ . dprobit college ability Iteration 0: log likelihood = -830.34831 Iteration 1: log likelihood = -652.46872 Iteration 2: log likelihood = -640.95584 Iteration 3: log likelihood = -640.78572 Iteration 4: log likelihood = -640.78568 ************************** MARGINAL EFFECTS FROM PROBIT MODEL ************************* Probit estimates Number of obs = 1217 LR chi2(1) = 379.13 Prob > chi2 = 0.0000 Log likelihood = -640.78568 Pseudo R2 = 0.2283 ------------------------------------------------------------------------------ college | dF/dx Std. Err. z P>|z| x-bar [ 95% C.I. ] ---------+-------------------------------------------------------------------- ability | .3397726 .0196691 16.72 0.000 -.000033 .301222 .378323 ---------+-------------------------------------------------------------------- obs. P | .4264585 pred. P | .386128 (at x-bar) ------------------------------------------------------------------------------ z and P>|z| are the test of the underlying coefficient being 0 . logit college ability Iteration 0: log likelihood = -830.34831 Iteration 1: log likelihood = -655.20523 Iteration 2: log likelihood = -638.96983 Iteration 3: log likelihood = -638.24636 Iteration 4: log likelihood = -638.24437 ***************************** LOGIT MODEL ESTIMATES **************************** Logit estimates Number of obs = 1217 LR chi2(1) = 384.21 Prob > chi2 = 0.0000 Log likelihood = -638.24437 Pseudo R2 = 0.2314 ------------------------------------------------------------------------------ college | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- ability | 1.544897 .1003651 15.39 0.000 1.348185 1.741609 _cons | -.5220004 .0730981 -7.14 0.000 -.6652701 -.3787306 ------------------------------------------------------------------------------ -------------------------------------------------------------------------------