You can download the data for this chapter by clicking here: https://stats.idre.ucla.edu/wp-content/uploads/2016/02/colvoucher.sas7bdat.
Descriptive statistics for Table 11.1 on page 270.
proc means data = colvoucher;
var finish8th use_fin_aid won_lottry base_age male;
run;
The MEANS Procedure
Variable Label N Mean Std Dev Minimum
finish8th finish8th 1171 0.6814688 0.4661058 0
use_fin_aid use_fin_aid 1171 0.5815542 0.4935148 0
won_lottry student won voucher 1171 0.5055508 0.5001828 0
base_age base_age 1171 12.0042699 1.3470381 7.0000000
male gender: equals 1 if applicant is male 1171 0.5046968 0.5001916 0
Variable Label Maximum
finish8th finish8th 1.0000000
use_fin_aid use_fin_aid 1.0000000
won_lottry student won voucher 1.0000000
base_age base_age 17.0000000
male gender: equals 1 if applicant is male 1.0000000
proc means data = colvoucher;
class won_lottry;
var finish8th use_fin_aid base_age male;
run;
The MEANS Procedure
student won N
voucher Obs Variable Label N Mean
0 579 finish8th finish8th 579 0.6252159
use_fin_aid use_fin_aid 579 0.2400691
base_age base_age 579 12.0362694
male gender: equals 1 if applicant is male 579 0.5043178
1 592 finish8th finish8th 592 0.7364865
use_fin_aid use_fin_aid 592 0.9155405
base_age base_age 592 11.9729730
male gender: equals 1 if applicant is male 592 0.5050676
student won N
voucher Obs Variable Label Std Dev Minimum
0 579 finish8th finish8th 0.4844857 0
use_fin_aid use_fin_aid 0.4274945 0
base_age base_age 1.3518143 7.0000000
male gender: equals 1 if applicant is male 0.5004137 0
1 592 finish8th finish8th 0.4409110 0
use_fin_aid use_fin_aid 0.2783108 0
base_age base_age 1.3427549 9.0000000
male gender: equals 1 if applicant is male 0.5003971 0
student won N
voucher Obs Variable Label Maximum
0 579 finish8th finish8th 1.0000000
use_fin_aid use_fin_aid 1.0000000
base_age base_age 16.0000000
male gender: equals 1 if applicant is male 1.0000000
1 592 finish8th finish8th 1.0000000
use_fin_aid use_fin_aid 1.0000000
base_age base_age 17.0000000
male gender: equals 1 if applicant is male 1.0000000
Significance tests for Table 11.1 on page 270.
proc glm data = colvoucher;
model finish8th = won_lottry;
run;
The GLM Procedure
Number of Observations Read 1171
Number of Observations Used 1171
The GLM Procedure
Dependent Variable: finish8th finish8th
Sum of
Source DF Squares Mean Square F Value Pr > F
Model 1 3.6241337 3.6241337 16.91 <.0001
Error 1169 250.5637399 0.2143402
Corrected Total 1170 254.1878736
R-Square Coeff Var Root MSE finish8th Mean
0.014258 67.93692 0.462969 0.681469
Source DF Type I SS Mean Square F Value Pr > F
won_lottry 1 3.62413371 3.62413371 16.91 <.0001
Source DF Type III SS Mean Square F Value Pr > F
won_lottry 1 3.62413371 3.62413371 16.91 <.0001
Standard
Parameter Estimate Error t Value Pr > |t|
Intercept 0.6252158895 0.01924033 32.50 <.0001
won_lottry 0.1112705970 0.02706015 4.11 <.0001
proc glm data = colvoucher;
model use_fin_aid = won_lottry;
run;
The GLM Procedure
Number of Observations Read 1171
Number of Observations Used 1171
The GLM Procedure
Dependent Variable: use_fin_aid use_fin_aid
Sum of
Source DF Squares Mean Square F Value Pr > F
Model 1 133.5541470 133.5541470 1031.16 <.0001
Error 1169 151.4074243 0.1295188
Corrected Total 1170 284.9615713
R-Square Coeff Var Root MSE use_fin_aid Mean
0.468674 61.88368 0.359887 0.581554
Source DF Type I SS Mean Square F Value Pr > F
won_lottry 1 133.5541470 133.5541470 1031.16 <.0001
Source DF Type III SS Mean Square F Value Pr > F
won_lottry 1 133.5541470 133.5541470 1031.16 <.0001
Standard
Parameter Estimate Error t Value Pr > |t|
Intercept 0.2400690846 0.01495640 16.05 <.0001
won_lottry 0.6754714559 0.02103510 32.11 <.0001
proc glm data = colvoucher;
model base_age = won_lottry;
run;
The GLM Procedure
Number of Observations Read 1171
Number of Observations Used 1171
The GLM Procedure
Dependent Variable: base_age base_age
Sum of
Source DF Squares Mean Square F Value Pr > F
Model 1 1.172741 1.172741 0.65 0.4217
Error 1169 2121.805910 1.815061
Corrected Total 1170 2122.978651
R-Square Coeff Var Root MSE base_age Mean
0.000552 11.22302 1.347242 12.00427
Source DF Type I SS Mean Square F Value Pr > F
won_lottry 1 1.17274119 1.17274119 0.65 0.4217
Source DF Type III SS Mean Square F Value Pr > F
won_lottry 1 1.17274119 1.17274119 0.65 0.4217
Standard
Parameter Estimate Error t Value Pr > |t|
Intercept 12.03626943 0.05598946 214.97 <.0001
won_lottry -0.06329646 0.07874516 -0.80 0.4217
proc glm data = colvoucher;
model male = won_lottry;
run;
The GLM Procedure
Number of Observations Read 1171
Number of Observations Used 1171
The GLM Procedure
Dependent Variable: male gender: equals 1 if applicant is male
Sum of
Source DF Squares Mean Square F Value Pr > F
Model 1 0.0001646 0.0001646 0.00 0.9796
Error 1169 292.7240028 0.2504055
Corrected Total 1170 292.7241674
R-Square Coeff Var Root MSE male Mean
0.000001 99.14968 0.500405 0.504697
Source DF Type I SS Mean Square F Value Pr > F
won_lottry 1 0.00016455 0.00016455 0.00 0.9796
Source DF Type III SS Mean Square F Value Pr > F
won_lottry 1 0.00016455 0.00016455 0.00 0.9796
Standard
Parameter Estimate Error t Value Pr > |t|
Intercept 0.5043177893 0.02079614 24.25 <.0001
won_lottry 0.0007497783 0.02924827 0.03 0.9796
Naive OLS estimate of the model shown in Table 11.2 on page 273.
proc glm data = colvoucher;
model finish8th = use_fin_aid base_Age male;
run;
The GLM Procedure
Number of Observations Read 1171
Number of Observations Used 1171
13:58 Monday, January 10, 2011 424
The GLM Procedure
Dependent Variable: finish8th finish8th
Sum of
Source DF Squares Mean Square F Value Pr > F
Model 3 16.2307260 5.4102420 26.53 <.0001
Error 1167 237.9571476 0.2039050
Corrected Total 1170 254.1878736
R-Square Coeff Var Root MSE finish8th Mean
0.063853 66.26252 0.451558 0.681469
Source DF Type I SS Mean Square F Value Pr > F
use_fin_aid 1 5.04596346 5.04596346 24.75 <.0001
base_age 1 9.04074811 9.04074811 44.34 <.0001
male 1 2.14401441 2.14401441 10.51 0.0012
Source DF Type III SS Mean Square F Value Pr > F
use_fin_aid 1 4.14983764 4.14983764 20.35 <.0001
base_age 1 8.33854172 8.33854172 40.89 <.0001
male 1 2.14401441 2.14401441 10.51 0.0012
Standard
Parameter Estimate Error t Value Pr > |t|
Intercept 1.410282818 0.12060311 11.69 <.0001
use_fin_aid 0.120908429 0.02680125 4.51 <.0001
base_age -0.062960958 0.00984556 -6.39 <.0001
male -0.085850473 0.02647542 -3.24 0.0012
Instrumental-variables (2SLS) from Table 11.2 on page 273.
proc syslin data = colvoucher 2sls first;
endogenous use_fin_aid;
instruments won_lottry base_age male;
model finish8th = use_fin_aid base_age male;
run;
The SYSLIN Procedure
First Stage Regression Statistics
Model First St
Dependent Variable use_fin_aid
Label use_fin_aid
Analysis of Variance
Sum of Mean
Source DF Squares Square F Value Pr > F
Model 3 134.1991 44.73303 346.26 <.0001
Error 1167 150.7625 0.129188
Corrected Total 1170 284.9616
Root MSE 0.35943 R-Square 0.47094
Dependent Mean 0.58155 Adj R-Sq 0.46958
Coeff Var 61.80463
Parameter Estimates
Parameter Standard Variable
Variable DF Estimate Error t Value Pr > |t| Label
Intercept 1 0.432760 0.095159 4.55 <.0001 Intercept
won_lottry 1 0.674527 0.021014 32.10 <.0001 student won voucher
base_age 1 -0.01516 0.007826 -1.94 0.0530 base_age
male 1 -0.02026 0.021070 -0.96 0.3365 gender: equals 1 if
applicant is male
Model First St
Dependent Variable finish8th
Label finish8th
Analysis of Variance
Sum of Mean
Source DF Squares Square F Value Pr > F
Model 3 15.44596 5.148654 25.17 <.0001
Error 1167 238.7419 0.204577
Corrected Total 1170 254.1879
The SYSLIN Procedure
First Stage Regression Statistics
Root MSE 0.45230 R-Square 0.06077
Dependent Mean 0.68147 Adj R-Sq 0.05835
Coeff Var 66.37170
Parameter Estimates
Parameter Standard Variable
Variable DF Estimate Error t Value Pr > |t| Label
Intercept 1 1.446936 0.119748 12.08 <.0001 Intercept
won_lottry 1 0.107250 0.026444 4.06 <.0001 student won voucher
base_age 1 -0.06457 0.009848 -6.56 <.0001 base_age
male 1 -0.08837 0.026514 -3.33 0.0009 gender: equals 1 if
applicant is male
The SYSLIN Procedure
Two-Stage Least Squares Estimation
Model finish8th
Dependent Variable finish8th
Label finish8th
Analysis of Variance
Sum of Mean
Source DF Squares Square F Value Pr > F
Model 3 15.44596 5.148654 25.21 <.0001
Error 1167 238.3690 0.204258
Corrected Total 1170 254.1879
Root MSE 0.45195 R-Square 0.06086
Dependent Mean 0.68147 Adj R-Sq 0.05844
Coeff Var 66.31984
Parameter Estimates
Parameter Standard Variable
Variable DF Estimate Error t Value Pr > |t| Label
Intercept 1 1.378128 0.123090 11.20 <.0001 Intercept
use_fin_aid 1 0.159000 0.039173 4.06 <.0001 use_fin_aid
base_age 1 -0.06216 0.009872 -6.30 <.0001 base_age
male 1 -0.08514 0.026504 -3.21 0.0014 gender: equals 1 if
applicant is male