The Evidence on Teaching

Does Head Start Improve Children's Life Chances? Evidence from a Regression Discontinuity Design

Ludwig, J., & Miller, D. L. · 2007

grade Bnatural-experimentindependentunreplicated
Sample
All US counties (mortality); 43 counties near the cutoff (NELS attainment)
Population
Counties just above/below the 1965 Head Start grant-writing assistance cutoff.
Design
RD on the county funding cutoff. Placebo checks pass (no discontinuity for causes not plausibly affected, or for cohorts never exposed).
Key findings
Independent quasi-experimental support for Head Start's attainment and health effects, with clean placebo tests. Like the other durable-effect findings, it comes from an era when the counterfactual was little or no preschool.
Counterfactual
Just-below-cutoff counties receiving 50-100% less Head Start funding.
Genetic confound
RD at a policy cutoff; county-level assignment is not family selection.
Replication notes
Attainment effects ~2x larger for Black children but with large standard errors. External validity to the modern era is limited.
DOI / URL
10.1162/qjec.122.1.159

Effects

OutcomeMetricValueMeasureTimingVsHorizonClass
Mortality, children 5-9, from Head-Start-susceptible causes% of control mean33-54% reduction (preferred specifications)standardizedgrade-3business-as-usualover-2yrhealth
High school completionpp+2 to +5ppmixedadultbusiness-as-usualadulthoodattainment
College attendancepp+3 to +6pp against a control mean of 23ppmixedadultbusiness-as-usualadulthoodattainment
Placebo: cohorts never exposed (ages 55+)discontinuitynone on any outcome — the key placebo passesmixedadultbusiness-as-usualadulthoodattainment

Cited by