123 children, randomly assigned to two groups, followed for forty years. The best-known evidence for the long-term effect of preschool — and its limits.
The study ran from 1962 to 1967 in Ypsilanti, Michigan. It was designed and directed by the psychologist David Weikart; the HighScope Educational Research Foundation holds the material today. 123 children from low-income families, assessed as at high risk of school failure, were randomly assigned: 58 to the programme group and 65 to the control group. The programme included weekly home visits intended to teach mothers to extend the preschool curriculum at home.
Outcomes recorded at age 40 (programme group vs control): high-school graduation 65% vs 45%; in employment 76% vs 62%; median annual earnings $20,800 vs $15,300; holding a savings account 76% vs 50%. On offending: arrested five or more times in their lifetime 36% vs 55%; sentenced to prison or jail 28% vs 52%.
Now the financial side — and here caution is required. By HighScope’s own calculation, every dollar invested returned $16.14. But James Heckman and colleagues, in a re-analysis published in the Journal of Public Economics in 2010, gave a considerably more conservative estimate: annual social rates of return generally between 7 and 10%. The difference comes from valuing crime costs differently, accounting for the deadweight loss of taxation, and acknowledging flaws in the randomisation.
That last point matters and should not be hidden. Researchers at the University of Chicago have documented compromises in the original randomisation procedure. Perry is therefore a randomised study, but not a clean RCT by modern standards, and the sample is small. The conclusion: the results carry weight, but they should not be presented as the only evidence.
