Skip to main content
. Author manuscript; available in PMC: 2012 Dec 17.
Published in final edited form as: Quant Econom. 2010;1(1):1–46. doi: 10.3982/qe8

Table 5.

Main outcomes: Males, part 1.a

Effect
 p-Values
Outcome Age Ctl.
Mean		 Uncond.b Cond.
(Full)c 		Cond.
(Part.)d 		Naïvee Full
Lin.f 		Partial
Lin.g 		Part. Lin.
(adj.)h 		Gender
D-in-Di 		Available
Observations
Education
   Mentally Impaired? ≤19 0.33 −0.13 −0.19 −0.17 0.106 0.072 0.057 0.190 0.337 66
   Yrs. in Disciplinary Program ≤19 0.42 −0.12 −0.26 −0.24 0.313 0.153 0.134 0.334 0.945 66
   Yrs. of Special Services ≤14 0.46 −0.04 −0.10 −0.09 0.458 0.256 0.205 0.349 0.153 72
   Learning Disabled? ≤19 0.08 0.08 0.08 0.07 0.840 0.841 0.766 0.766 0.029 66
   Highest Grade Completed 19 11.28 0.08 −0.01 0.15 0.429 0.383 0.312 0.718 0.052 72
   Grade Point Average 19 1.79 0.02 −0.01 0.07 0.464 0.517 0.333 0.716 0.009 47
   Vocational Training Certificate ≤40 0.33 0.06 0.06 0.03 0.231 0.304 0.406 0.729 0.500 72
   High School Graduation 19 0.51 −0.03 0.00 0.02 0.633 0.510 0.416 0.583 0.003 72
   # Years Held Back ≤19 0.39 0.08 0.12 0.09 0.740 0.852 0.745 0.745 0.106 66
Health
   Alive 40 0.92 0.05 0.05 0.06 0.160 0.174 0.146 0.604 0.909 72
   No Sick Days in Bed, Past Yr. 27 0.38 0.10 0.14 0.12 0.208 0.135 0.162 0.582 0.412 70
   No Treat. for Illness, Past 5 Yrs. 27 0.64 0.00 0.01 0.03 0.465 0.417 0.375 0.826 0.806 70
   No Doctors for Illness, Past Yr. 19 0.56 0.07 0.02 0.02 0.210 0.435 0.453 0.835 0.609 72
   No Non-Routine Care, Past Yr. 27 0.17 −0.03 −0.02 −0.01 0.600 0.548 0.548 0.823 0.549 63
   No Health Problems 19 0.95 −0.07 −0.08 −0.08 0.849 0.843 0.862 0.862 0.308 72
   Infrequent Alcohol Use 27 0.58 0.18 0.21 0.20 0.072 0.024 0.052 0.139 0.924 66
   No Tobacco Use 27 0.46 0.12 0.10 0.09 0.143 0.220 0.260 0.436 0.965 70
   Routine Annual Health Exam 27 0.74 −0.04 0.01 0.01 0.622 0.397 0.451 0.451 0.867 68
Crime
   # Non-Juv. Arrests ≤27 5.36 −2.33 −2.64 −2.71 0.029 0.028 0.017 0.024 0.571 72
   # Fel. Arrests ≤27 2.33 −1.12 −1.07 −1.15 0.046 0.081 0.043 0.101 — 72
   Any Non-Juv. Arrests ≤27 0.72 −0.02 −0.05 −0.05 0.501 0.422 0.291 0.418 0.440 72
   Any Fel. Arrests ≤27 0.49 0.00 −0.01 −0.01 0.494 0.575 0.442 0.442 — 72
   Any Non-Juv. Arrests ≤40 0.92 −0.14 −0.12 −0.12 0.090 0.124 0.078 0.192 0.463 72
   Any Fel. Arrests ≤40 0.44 −0.16 −0.15 −0.16 0.047 0.133 0.083 0.191 — 72
   Any Arrests ≤40 0.95 −0.13 −0.11 −0.09 0.072 0.142 0.123 0.181 0.824 72
   Any Misd. Arrests ≤40 0.87 −0.11 −0.08 −0.07 0.166 0.281 0.191 0.191 0.519 72
   # Misd. Arrests ≤40 8.46 −3.13 −3.42 −3.64 0.037 0.043 0.021 0.039 0.549 72
   # Non-Juv. Arrests ≤40 11.72 −4.26 −4.45 −4.85 0.039 0.053 0.025 0.041 0.458 72
   # Total Arrests ≤40 12.41 −4.20 −4.44 −4.88 0.056 0.073 0.036 0.053 0.566 72
   # Fel. Arrests ≤40 3.26 −1.14 −1.03 −1.20 0.112 0.173 0.092 0.092 — 72
   # Non-Victimless Chargesj ≤40 3.08 1.59 1.65 1.65 0.029 0.048 0.027 0.061 0.175 72
   # Total Charges ≤40 13.38 4.38 5.08 5.08 0.063 0.081 0.041 0.075 0.637 72
   Total Crime Costk ≤40 775.90 −351.22 −515.10 −515.10 0.153 0.108 0.070 0.070 0.858 72
   Any Non-Victimless Chargesj ≤40 0.62 0.16 0.15 0.15 0.105 0.179 0.112 0.259 0.957 72
   Ever Incarcerated ≤40 0.23 −0.08 −0.11 −0.12 0.260 0.159 0.114 0.202 0.563 72
   Any Charges ≤40 0.95 0.13 0.09 0.09 0.072 0.142 0.125 0.125 0.799 72
a

Monetary values adjusted to thousands of year-2006 dollars using annual national CPI. p-values below 0.1 are in bold.

b

Unconditional difference in means between the treatment and control groups.

c

Conditional treatment effect with linear covariates Stanford–Binet IQ, Socioeconomic Status index (SES), maternal employment, father’s presence at study entry—this is also the effect for the Freedman–Lane procedure under a full linearity assumption, whose respective p-value is computed in column “Full Lin.”

d

Conditional treatment effect as in the previous column except that SES is replaced with an indicator for SES above/below the median, so that the corresponding p-value is computed in the column “Partial Lin.” This specification generates p-values used in the stepdown procedure.

e

One-sided p-values for the hypothesis of no treatment effect based on conditional permutation inference, without orbit restrictions or linear covariates—estimated effect size in the “Uncond.” column.

f

One-sided p-values for the hypothesis of no treatment effect based on the Freedman–Lane procedure, without restricting permutation orbits and assuming linearity in all covariates (maternal employment, paternal presence, Socioeconomic Status index (SES), and Stanford–Binet IQ)—estimated effect size in the “conditional effect” column.

g

One-sided p-values for the hypothesis of no treatment effect based on the Freedman–Lane procedure, using the linear covariates maternal employment, paternal presence, and Stanford–Binet IQ, and restricting permutation orbits within strata formed by Socioeconomic Status index (SES) being above or below the sample median and permuting siblings as a block.

h

p-values from the previous column, adjusted for multiple inference using the stepdown procedure.

i

Two-sided p-value for the null hypothesis of no gender difference in mean treatment effects, tested using mean differences between treatments and controls using the conditioning and orbit restriction setup described in footnote f.

j

Non-victimless crimes are those associated with victimization costs: murder, rape, robbery, assault, burglary, larceny, and motor vehicle theft (see Heckman et al. (2010a) for details)

k

Total crime costs include victimization, police, justice, and incarceration costs, where victimizations are estimated from arrest records for each type of crime using data from urban areas of the Midwest, police and court costs are based on historical Michigan unit costs, and the victimization cost of fatal crime takes into account the statistical value of life (see Heckman et al. (2010a) for details).