As I described last week, the Department of Health and Human Services has been sitting on an evaluation of the Head Start government run pre-school program. Well, today the study was released (and it’s not even a Friday!). (Update: HHS moved the link to the study to here.)
As the leaks suggested, the study found virtually no lasting effects to participation in Head Start. The study used a gold-standard, random assignment design and had a very large nationally representative sample. This was a well done study (even if it mysteriously took more than 3 years after data collection was complete to release the results).
For students who were randomly assigned to Head Start or not at the age of 4, the researchers collected 19 measures of cognitive impacts at the end of kindergarten and 22 measures when those students finished 1st grade. Of those 41 measures only 1 was significant and positive. The remaining 40 showed no statistically significant difference. The one significant effect was for receptive vocabulary, which showed no significant advantage for Head Start students after kindergarten but somehow re-emerged at the end of 1st grade.
The study used the more relaxed p< .1 standard for statistical significance, so we could have seen about 4 significant differences by chance alone and only saw 1. That positive effect had an effect size of .09, which is relatively modest.
For students randomly assigned to Head Start or not at the age of 3, the researchers also collected 41 measures of lasting cognitive effects. This time they found 2 statistically significant positive effects and 1 statistically significant negative effect. For the students who began at age 3 they showed a .08 effect size benefit from Head Start in oral comprehension after first grade and a .26 effect size benefit in spanish vocabulary after kindergarten but a .19 effect size decline in math ability at the end of kindergarten. Again, 38 of the 41 measures of lasting effects showed no difference and the few significant effects (which could be produced by chance) showed mixed results.
I think it is safe to say from this very rigorous evaluation that Head Start had no lasting effect on the academic preparation of students.
The study also measured lasting effects on student behavior and emotion as well as the skills of parents. Again, the effects were largely null and the few significant differences were in mixed directions. The few positive effects from these categories were from parent reports and the few negative tended to come from teacher reports.
The long and short of it is that the government has a giant and enormously expensive pre-school program that has made basically no difference for the students who participate in it. And folks are proposing that we expand government pre-school to include all students. Those same folks have some bridges they’d like to sell.
(edited for clarity)
Posted by Jay P. Greene 
Jay P. Greene's Blog 