Perry Preschool treated only 58 kids with small-sample biases. Abecedarian treated 57 infants with evaluations by biased staff. IHDP large replication showed no effects at age 18 on cognition, achievement, crime or dropout. Head Start rigorous evaluation found only 2 of 94 outcomes significant after corrections. Head Start had 15-20% no-shows indicating selection bias. Scalable programs like Head Start show no third-grade gains persisting. Best programs use unreplicable high costs and motivated staff. No evidence disadvantaged subgroups benefited in IHDP. Early gains in small studies fade by school age. Government lacks methods to help non-flourishing kids via programs. Too many kids born to uncaring incompetent parents need better origins not interventions.