National Job Corps Study
Mathematica Policy Research (Schochet, Burghardt, Glazerman) · United States (national) · 1995
Summary
The National Job Corps Study produced a paradox: one of the largest and most rigorous evaluations of a US workforce program found that the most expensive per-participant workforce program ($16,500 per enrollee in 1995 dollars) produced modest short-term earnings gains that faded by year 4. The $22/week earnings advantage in year 4 was real and statistically significant, but the implied internal rate of return was low. Yet the program also produced reductions in criminal arrests during enrollment — a benefit not captured in earnings-only analyses. The study's most important contribution may have been methodological: demonstrating that intensive residential training was randomizable, and that long-run follow-up was essential for any honest program evaluation. The mixed finding has sustained debate about Job Corps for three decades, with supporters emphasizing subgroup effects (older youth, those with criminal histories) and critics emphasizing the aggregate cost-effectiveness.
Research question
"Does intensive residential vocational training for at-risk youth aged 16–24 improve earnings and reduce criminal behavior?"
Methodology
Intervention
Job Corps is a federally funded residential education and training program serving low-income youth aged 16–24. Participants receive vocational training, high school diploma or GED preparation, social skills instruction, health care, and housing at one of 125 residential centers across the US. The program lasts an average of 8 months.
Assignment
Randomized controlled trial; applicants who passed initial eligibility screening were randomly assigned to be allowed to enroll (treatment) or denied enrollment for 3 years (control); one of the largest social policy RCTs ever conducted
Sample size
Approximately 15,400 youth (9,409 treatment, 5,977 control) at 80 study sites
Primary outcome
Earnings and employment; GED and vocational credential attainment; criminal behavior and arrest rates
Effect estimate
At 4 years post-assignment: 12% increase in earnings ($22/week); 1.6 percentage point increase in employment; significant reduction in criminal behavior during the program years; criminal behavior effects faded after the program ended; credential attainment significantly higher in treatment group
Decision
National Job Corps Study remains the largest RCT ever conducted on a US workforce program; findings showed positive but modest labor market effects that fade over time; subsequent cost-benefit analyses found positive returns for some subgroups (particularly older youth and those with prior criminal records); program continues to operate at $1.7 billion annually; study design became a model for subsequent workforce evaluation
Result
Mixed
At 4 years post-assignment: 12% increase in earnings ($22/week); 1.6 percentage point increase in employment; significant reduction in criminal behavior during the program years; criminal behavior effects faded after the program ended; credential attainment significantly higher in treatment group
Evidence strength
Moderate
Randomized trial; replication status unknown or limited.
Replication status
N/A
Institution
Mathematica Policy Research (Schochet, Burghardt, Glazerman)
Location
United States (national)
Year
1995
Policy area
Education
Mechanism
Human capital