As the evidence has piled up that early childhood education programs pay off in the long run, Democratic presidential candidates have vowed to expand them. 

“Study after study has shown that regular access to high-quality child care promotes literacy skills, cognitive development, and healthy behaviors,” wrote Elizabeth Warren in February. “These are long-term benefits: quality early education produces better health, educational, and employment outcomes well into adulthood.

Indeed, a raft of research has shown that children see benefits even many years later from attending early childhood programs like the Perry preschool initiative and federally funded Head Start. A prominent 2009 study by Harvard economist David Deming showed that children who attended Head Start between in the 1970s and ‘80s were 8.5 percentage points more likely to graduate high school as a result. 

That’s why a new study comes a surprise. When researchers used similar methods as Deming to look at students born later, mostly between 1986 and 1996, they found that Head Start provided no clear long-term benefits. If anything, the program led to somewhat worse outcomes. 

The research offers some of the most up-to-date information about how Head Start affects students years later, and raises questions whether Head Start continues to deliver the positive outcomes it did in the past. In 2016-17, a third of U.S. 3- to 5-year-olds living in poverty were enrolled in the program.

“We’re using methods that have been accepted for a long time … and now we’ve got a bunch of negative results,” said Dylan Lukes, a Harvard graduate student and one of the authors. 

The researchers say the results should be interpreted with caution, as it’s a single study that goes against the grain of other research. A spokesperson for the division of the U.S. Department of Health and Human Services that administers Head Start emphasized that point, too. 

The agency “takes into account the large body of research and evidence on Head Start when developing Head Start program policy and practice, rather than considering any single study in isolation,” said Patrick Fisher. “Indeed, the study authors cite this broader body of evidence, which has demonstrated a variety of positive long-term outcomes for Head Start participants.” 

Still, it’s not clear what explains the new findings. 

“I think the paper is well-executed, and I believe the results,” said Deming. “The key question is how to interpret them.”

More recent data finds no long-term benefits of Head Start. Why?

Head Start launched in 1965 as part of President Lyndon Johnson’s “war on poverty” and was designed to prepare young children from low-income families for kindergarten with early education, health services, and family support. Today, over 1 million children, largely 3- and 4-year-olds, participate in the federally funded program.

Deming’s study compared Head Start children to their siblings who didn’t go to preschool. The Head Start attendees saw higher high school graduation rates but also improvements in college attendance and health. Other studies, using different methods, have found similarly encouraging results

Another recent paper showed that the effects spanned generations: even the children of children who attended Head Start soon after it began were more likely to graduate high school and less likely to be involved with the criminal justice system. 

But when Lukes (a student of Deming’s) and researchers at the University of California, Irvine, updated Deming’s study with more recent data, they found something puzzling. 

For those born between 1976 and 1986, they, like Deming, found Head Start provided clear benefits. For children born later, though — largely between 1986 and 1996 — the gains disappeared and even reversed in some cases. As young adults, those Head Start participants were 8 percentage points more likely to be unemployed or not in school. 

When the researchers combined both groups, it was a wash — no clear long-term effects, good or bad. 

It’s not clear what explains the strikingly different findings across groups of students. Other research has shown that the benefits of Head Start are larger when children also attend well-funded K-12 schools with better teachers. But school funding has generally risen over time, so declining K-12 school quality seems an unlikely explanation. 

Deming hypothesizes that conditions for low-income families have improved since earlier research on Head Start. The program doesn’t just provide education, but also a suite of health and anti-poverty services. Since later groups of students were born to mothers who were more educated and into families with somewhat higher incomes, perhaps some of those Head Start services were less necessary.

“Even though in a narrow sense the impact of Head Start has diminished, in a broader sense it is a huge victory that more and more children are getting high-quality physical, mental, and academic support in the critical early years,” said Deming. “The study doesn’t show this directly, but it fits with the narrative.” 

Overall, research on pre-K is still largely positive

It can be hard to make sense of the competing research on early childhood education. One reason is that by the time researchers can see the long-run effects, the program in question is decades old.

That means it’s difficult to know what the latest results tell us about today’s early childhood initiatives.

The latest research did find an important pattern: the kids who were in Head Start between 1976 and 1986 not only saw long-run benefits, they also saw more immediate improvements in math and reading scores while in school compared to students who didn’t attend pre-K at all. The ‘86-’96 students didn’t see those test-score gains.

That’s evidence that policymakers should be reasonably optimistic about early childhood programs that produce immediate benefits, but cautious about programs that don’t. 

More recent research on pre-K, including Head Start, usually finds immediate benefits — though they may dissipate as students move through school. Complicating matters further: the benefits sometimes reappear.

That all means that there’s still a strong research-based case for preschool programs.

“The mixed evidence in this paper suggests that it is possible that Head Start has been less effective at improving outcomes for more recent participants,” said Chloe Gibbs of the University of Notre Dame, who has conducted a number of studies on Head Start. “But there is still a large body of evidence documenting Head Start’s important long-term effects.” 

And the researchers in the latest study say that they want their findings to be replicated using different approaches before any policy decisions are made based on the results.

“It would be way too premature to make any sort of policy changes off one paper,” said Lukes.