Early childhood intervention is particularly cost-beneficial when it reduces justice involvement, but ingredients that contribute to this outcome are unknown. The goal of this study was to estimate the effects of two common early childhood intervention ingredients--home visits and center-based education--on juvenile justice involvement. The Infant Health and Development Program (IHDP) randomized 1090 premature and low-birth-weight babies to intervention or control groups. Intervention group families were offered home visits from birth to age 3 years and high-quality center-based early childhood education from ages 1 to 3 years, but varied in their take-up of each intervention component. We estimated (1) intent-to-treat effects and (2) the effects of families' level of participation in each intervention component, using a novel stratification approach to minimize the impact of self-selection bias on dosage. Outcomes were children's risk of being stopped by police, arrested, or incarcerated, by age 18 years. Intent-to-treat analyses showed no effects of the IHDP for both sexes combined, nor for girls only, on any of the three outcomes, but there was an intent-to-treat effect on boys' risk of being arrested, OR = 0.43 (95% CI 0.24, 0.76). Analyses of dosage effects showed that, for both sexes combined, participation in the center-based educational component decreased the odds of being stopped by the police by 3% for each month of services. For boys only, the odds of being arrested decreased by 4% with each month of home visits and by 4% with each month of center-based educational services. We conclude that high-quality center-based early childhood education and, to some extent, home visits, reduce justice involvement among biologically vulnerable children, especially boys.