Almost since Head Start’s creation half a century ago as part of the War on Poverty, advocates have sought a research-based answer to this question: Does it work?
Two new studies examining the long-term impact of the $8.6 billion federal preschool program offer support to those who say the answer is yes.
Those studies—one of which took a look at Head Start participants in middle school and another that examined life outcomes at adulthood—say that Head Start has measurable long-term benefits for certain outcomes. Those include higher levels of high school and college completion found in one study, and lower levels of chronic absenteeism and grade retention found in another.
Whether Head Start can work “is an antiquated question at this point,” said Deborah Phillips, a professor of psychology at Georgetown University and the co-author of a study that looked at middle school outcomes for children who took part in a Head Start program in Tulsa, Okla.
“The answer is, yes it can,” Phillips said. “The question now is, under what circumstances does the Head Start program produce these effects?”
At the same time, these studies raise other questions: While both showed some positive effects for students overall, that wasn’t always the case when looking at some subgroups, such as black students or boys. And both studies acknowledge that there might be unknowable differences between the Head Start students and the comparison group that might affect the magnitude of the results, though they attempted to control for these factors.
Addressing ‘Fade-Out’
The most common critique of Head Start is that its academic benefits fade out by the time children are a few years into elementary school.
The touchstone study in this area was the federally funded National Head Start Impact Study, which followed a group of students who enrolled in Head Start and compared them with another group of students who did not enroll because the program in their area was oversubscribed. (Some of those children did go on to other preschools, or to other Head Start programs that had room for them.)
That study found initial positive impacts but also found that by the end of 3rd grade, Head Start students were performing very similarly to their peers. No other Head Start research has matched the large-scale, randomized experimental design of the impact study.
But supporters of Head Start have argued that the program offers beneficial “sleeper effects” that are best measured later in a participant’s life. These two new studies, while not having the same “gold standard” research design as the federal study, still offer important information, their authors say.
The Tulsa research followed children who enrolled in the Head Start program that is run by the Community Action Project of Tulsa. It found that 7th graders who had attended that program had measurably higher math scores than students who had not enrolled in Head Start or in another preschool program that was run by the Tulsa district.
The Tulsa research also found that the former Head Start students were less likely to be retained prior to 8th grade: 14 percent of them had been retained, compared with 20 percent of the students in the control group.
And while 9 percent of the control group had been absent more than 18 days in the school year, that was true for only 6 percent of the former Head Start students.
Phillips said the Tulsa research shows that some academic effects are still measurable, along with nonacademic effects that are important, such as showing up for school regularly.
But when breaking out results into subgroups, the Tulsa study did not find any statistically significant Head Start effects on boys and on African-American children. The measurable positive effects were found for girls, Hispanic students, white students, and students who were eligible for free lunches.
“This caught us by surprise,” Phillips said. “Unfortunately, we do not have a lot of data about what happens in the schools. These kids got a strong boost into school, but something is happening between 3rd and 8th grade that is not supportive of African-American kids and boys.” That finding merits deeper research, she said.
It’s also noteworthy that CAP Tulsa in 2005-06, when these children were enrolled, was different from other Head Start programs. Although Head Start is federally funded, the government does not directly run the programs. Instead, public agencies and community groups such as CAP Tulsa provide space, personnel, materials, and professional development for the Head Start programs they manage.
Oklahoma has a universal pre-K program, and CAP Tulsa meets many of the same requirements that the school-run programs do. The lead teachers all have bachelor’s degrees and special training in early childhood, which was not true of all Head Start programs in 2005-06. The teachers are also paid wages that are comparable to public school teachers. The program also scored high on measures of academic instruction and instructional quality. That means the effects seen here might not generalize to Head Start nationally.
But Head Start has undertaken a variety of different quality-improvement measures, Phillips said. Lead teachers are now required to have a bachelor’s degree, and low-rated programs have been forced to compete for renewed funding.
“What’s very encouraging today is that the national Head Start program is moving in the direction of what’s going on in Tulsa,” Phillips said.
Impacts on Adulthood
Another study of Head Start effects goes much further than middle school. The Hamilton Project, an economic-policy initiative of the Brookings Institution, used the National Longitudinal Survey of Youth, which measures a wide variety of economic and social factors, including whether participants enrolled in Head Start.
The researchers compared Head Start participants with their siblings who either attended another Head Start program or did not go to preschool.
That study found that participating in Head Start increased a person’s likelihood of graduating from high school, pursuing post-high-school education, and completing a post-high-school program, which might include a certificate, associate degree, or bachelor’s degree.
Another result from the long-term study was that some children who took part in Head Start grew up into adults who invested more in positive parenting practices, such as reading to their child or showing physical affection. That is in comparison with their siblings who attended another preschool program.
“This is the first of its kind in terms of a finding, so there’s much more to explore here. But it gives us some hope that this will continue to spill over into the next generation,” said Diane Whitmore Schanzenbach, a co-author of the research as well as the director of the Hamilton Project.
When broken out into subgroups, however, the parenting findings were only significant for former Head Start students who were black. For Hispanic and white participants, the effects were not statistically different from those of their siblings who attended a different preschool.
Grover J. “Russ” Whitehurst, the director of the Brown Center on Education Policy at Brookings, said that these studies don’t offer proof of Head Start’s effectiveness. For the longer-range study, that’s because, despite attempts at statistical controls, the studies cannot measure all the factors that might prompt a parent to enroll one child in Head Start and not a sibling. Or, in the case of the Tulsa research, there’s no way to determine why one family might seek out Head Start while another family might look for another type of early-care arrangement.
“I don’t think either of these two studies adds credible information to what we know about the impact of Head Start,” Whitehurst said. “The parent was observing something about the child, or observing something about her own circumstances that year. That is almost surely correlated with the measures of later outcomes.”
Schanzenbach said that the sibling-comparison method used in the Hamilton Project study, which also controlled for a number of additional child, maternal, and household factors, can produce important findings.
“What we want to know is what are the long-term impacts of preschool, and we’re not going to know that from an experiment unless we have a time machine,” Schanzenbach said. But this long-term-evaluation approach “adds a nice new dimension to the broader literature on this really important topic.”
