While American children generally acquire basic math and reading skills, they often fail to master more advanced skills like fractions, algebra or critical reading. (iStock)

Drew Bailey is an assistant professor and Greg Duncan is a professor in the School of Education at the University of California at Irvine. Candice Odgers is a professor of public policy, psychology and neuroscience at Duke University and a Jacobs Foundation research fellow.

States and the federal government spend more than $15 billion a year on preschool education. With that hefty price tag, we want early-childhood programs to work. And to reduce long-standing educational inequalities, we need them to work. So it’s encouraging when studies show that these types of interventions can give children a boost by the time they enter kindergarten.

Unfortunately, our investments in many early-childhood programs may be based on an inflated sense of their promise. Even our best efforts often produce only ephemeral gains.

We reviewed data from 67 high-quality interventions — all of which included some degree of pre-literacy and early math skill-building and most of which targeted economically disadvantaged children — and we found that the effects faded startlingly fast: falling by half within a year and by half again two years later.

Head Start — perhaps the best-known early-childhood program — was no exception. A recent national evaluation showed that a year of Head Start produced tangible benefits. For example, Head Start 4-year-olds scored significantly higher on three of the most widely used literacy assessments than did similarly disadvantaged children who had been placed on Head Start waitlists. The benefits were enough to cut in half the gap in pre-reading skills between Head Start students and the national average for children of the same age. When tested a year later, at the end of kindergarten, however, the reading skills of Head Start children were indistinguishable from those of comparison-group children.

We saw this pattern for one intervention after another.

Why is fadeout so common? It’s obvious that basic math and reading skills are necessary for acquiring more advanced math and reading skills; you need to be able to count to learn to add, and you need to understand the correspondence between letters and sounds to learn to read. And many studies have documented strong correlations between school-entry skills and later math and reading performance. So why don’t boosts in these early foundational skills lead to permanent advantages in the academic trajectories of disadvantaged kids?

Our work suggests that much of what children learn in early-childhood intervention programs are skills that kids typically pick up in kindergarten or first grade anyway. Fadeout is really a process of other children catching up — learning their letters, learning to count, learning to control their emotions and impulses.

What holds disadvantaged kids back throughout their schooling is not a failure to master the basics.

More fundamental to achievement are hard-to-change characteristics such as intelligence and conscientiousness, as well as persistent environmental factors that are difficult to change with a one-time educational intervention. American children raised in poverty must wade through a stream of stresses. They may have to move often or even be homeless for a time. They may be exposed to violence in their homes or neighborhoods. Their parents, dealing with stresses in their own lives, may not be able to provide as stimulating an upbringing as more economically comfortable parents can — leading to problems such as the 30 million-word gap. When children grow up in challenging home environments and move through mediocre classrooms, it’s not surprising that they are unable to translate early gains in basic math and reading skills into mastery of more advanced skills such as fractions, algebra or critical reading.

That’s not an argument for giving up on early-childhood interventions. There is evidence that pre-K helps ease the transition to full-day school, sustaining kids through a period of high vulnerability. It may reduce the need for special education placement or for kids to be held back a grade, which may in turn have positive cascade effects. Additionally, even if the initial benefits fade rapidly, a handful of the most rigorously implemented and evaluated programs appear to produce benefits that persist well into adulthood: Participants graduate from high school at a higher rate, earn more, have fewer run-ins with the law and lead healthier lives than peers from similar backgrounds. We don’t know exactly why.

It’s a mistake to assess these programs — and justify policy — by looking solely at metrics immediately after the conclusion of an intervention. It’s vital that we continue to track the longer-run effects.

But it’s also a mistake to expect that initial gains will sustain themselves. For lasting effects, we need to focus on skills that wouldn’t otherwise develop, do more to change a child’s environment and provide ongoing support, especially during sensitive periods of development such as early adolescence.

One intriguing set of interventions seeks to encourage children’s beliefs about their learning potential, help them affirm their values and enable them to see the point of their education. For example, asking minority middle school students to write about something they value (e.g., friendships or music) has been shown in some experiments to improve their GPA relative to their white peers for years afterward. Such interventions are attractive because they are inexpensive and can be tailored to different contexts, though there are questions about how reliably the results can be replicated in larger samples.

Another worthy model for interventions targets not just children but their caregivers, with the idea that improving parent-child interactions can affect the whole course of a child’s development. For example, low-income mothers randomly assigned to the Nurse-Family Partnership, which involves home visits during pregnancy and a child’s infancy, were subsequently less reliant on social services, and their children had better grade-point averages and achievement test scores in math and reading through elementary school.

A further way to improve children’s environments would be to offer more large-scale interventions. Recent research on North Carolina’s statewide More at Four program suggests that filling early-grade classrooms with enough students who have already mastered the basics may allow teachers to offer a more advanced curriculum from the start.

We need to focus on older children as well. Intensive interventions can help with mastery of advanced skills. A high school program offering a double dose of algebra classes has been shown to boost graduation rates. And students who attend career academies, which restructure large high schools into small learning communities and promote concrete vocational skills, can typically expect higher earnings as adults. Vocational skills, advanced math and literacy skills, and achievement-related beliefs and motivations are examples of what we call “trifecta skills” — skills that are malleable, fundamental to success and unlikely to develop in the absence of an intervention.

Early-childhood programs have the potential to improve the lives of millions of children. Everyone agrees that these programs have the best of intentions. But good intentions are not enough. We need to design interventions that generate persistent advantages for our children.