自变量：trans_1
因变量：lnfeducation
协变量：gender married work education_f education_m lnfincome lne_mail fasset
代码：egen trans1_p25=pctile(trans1), p(25)
egen trans1_p50=pctile(trans1), p(50)
egen trans1_p75=pctile(trans1), p(75)
generate cut=trans1_p25 if trans1<=trans1_p25
replace cut=trans1_p50 if trans1>trans1_p25 & trans1<=trans1_p50
replace cut=trans1_p75 if trans1>trans1_p50 & trans1<=trans1_p75
replace cut=0.0770572 if trans1>trans1_p75
matrix tp=(10\20\30\40\50\60\70\80\90\100)
doseresponse2 gender married work education_f education_m lnfincome lne_mail fasset, outcome(lnfeducation) t(trans1) gpscore(gps) predict(t_hat) sigma(sd) cutpoints(cut) index(mean) nq_gps(4) t_transf(ln) family(bin) link(logit) dose_response(does) tpoints(tp) delta(1) reg_type_t(cubic) reg_type_gps(cubic) interaction(1) bootstrap(yes) boot_reps(200) analysis(yes) filename("gpsm1") graph("gpsm1") detail

结果：
********************************************
ESTIMATE OF THE GENERALIZED PROPENSITY SCORE
********************************************

Generalized Propensity Score

******************************************************
Algorithm to estimate the generalized propensity score
******************************************************



Estimation of the propensity score

The log transformation of the treatment variable trans1 is used

                              T
-------------------------------------------------------------
      Percentiles      Smallest
1%    -4.906392      -4.906392
5%    -4.897616      -4.906392
10%    -4.897616      -4.906392       Obs               9,297
25%    -4.490526      -4.906392       Sum of wgt.       9,297

50%    -3.668333                      Mean           -3.84712
                        Largest       Std. dev.      .6275509
75%    -3.386563      -2.563208
90%    -3.223094      -2.563208       Variance       .3938201
95%    -3.130241      -2.563208       Skewness      -.4166525
99%    -2.563208      -2.563208       Kurtosis       1.968049
dependent variable T has negative values
r(499);

end of do-file

r(499);