?****************************************************** ? Problem Set 4 ? Tutorial in Microeconometrics summer term 2008 ? Katrin Sommerfeld ?****************************************************** options crt; freq N; read (file='L:\Microeconometrics\schweiz.raw') emp NLI age h educ exper LMS mar kid treat ; ?****************************************************************************** ? 1. Descriptive Statistics: msd(terse) @all; title 'Employment of women without tertiary education:'; select treat=0; msd(terse) emp; set emp0=@mean; select 1; title 'Employment of women with tertiary education'; select treat=1; msd(terse) emp; set emp1=@mean; select 1; print emp0 emp1; ?***************************************************************************** ? 2. Treatment effect in case of no difference between treatment and control group set ATE=emp1-emp0; print ATE; ?***************************************************************************** ? 3. Test for difference between treatment and control group title 'Women without tertiary education:'; select treat=0; msd @all; set nli0=@mean(2); set var0=@var(2); select 1; title 'Women with tertiary education'; select treat=1; msd @all; set nli1=@mean(2); set var1=@var(2); select 1; set t=(nli1-nli0)/(sqrt(var1+var0)); print t; ?***************************************************************************** ? 4. Treatment effects ? a) Probit estimation probit emp c NLI h exper LMS mar kid treat; ? b) Generate the individual probability of being employed in case of treatment or no treatment genr prob0 = cnorm(@coef(1) + NLI*@coef(2) + h*@coef(3) + exper*@coef(4) + LMS*@coef(5) + mar*@coef(6) + kid*@coef(7) + 0*@coef(8) ) ; genr prob1 = cnorm(@coef(1) + NLI*@coef(2) + h*@coef(3) + exper*@coef(4) + LMS*@coef(5) + mar*@coef(6) + kid*@coef(7) + 1*@coef(8) ) ; msd (terse) prob0 prob1; ? c) Generate the difference genr diff=prob1-prob0; title 'AVERAGE TREATMENT EFFECT = ATE'; msd(terse) diff; ? stop; ?***************************************************************************** title '5. AVERAGE TREATMENT EFFECT ON THE TREATED = ATT'; select treat=1; msd(terse) diff; select 1; end;