On February 26, the Wall Street Journal published an editorial criticizing President Obama’s proposal to expand preschool. The editorial was entitled “Head Start for All: Universal preschool and a government that won’t admit failure”. Given the prominence of the Wall Street Journal, this editorial has been widely circulated and cited by critics of proposals for expanding preschool. In response to several reader requests, I am examining some of the claims made in the editorial.
To begin with, the editorial’s title, “Head Start for All”, does not appear to be an accurate description of President Obama’s proposal. According to a White House fact sheet, President Obama’s preschool proposal would do the following:
“The President’s proposal will improve quality and expand access to preschool, through a cost sharing partnership with all 50 states, to extend federal funds to expand high-quality public preschool to reach all low- and moderate-income four-year olds from families at or below 200% of poverty. The U.S. Department of Education will allocate dollars to states based their share of four-year olds from low- and moderate-income families and funds would be distributed to local school districts and other partner providers to implement the program. The proposal would include an incentive for states to broaden participation in their public preschool program for additional middle-class families, which states may choose to reach and serve in a variety of ways, such as a sliding-scale arrangement….”
“Under the President’s proposal, investment in the federal Head Start program will continue to grow. The President’s plan will maintain and build on current Head Start investments, to support a greater share of infants, toddlers, and three-year olds in America’s Head Start centers, while state preschool settings will serve a greater share of four-year olds. “
Therefore, this proposal differs from “Head Start for All” in several respects:
- The proposal is largely funding expansion of state programs, not an expansion of directly federally-funded local Head Start centers.
- Under the proposal, many children would actually shift from Head Start centers at age 4 to state-designed programs to serve four-year-olds. It seems likely that there would be much more state and local flexibility under this program than under Head Start, although how much flexibility depends upon yet-to-be-released details in the proposal.
- The bulk of the federal funding goes to four-year olds in families below 200% of the poverty line, although it sounds as if some funds may encourage expansion of such programs with some fees to more middle-class families. While the proposal can be seen as expanding targeted preschool to more kids in a dramatic way with federal dollars, the degree to which the program advances truly universal pre-K will largely rest on state initiative and state dollars.
The editorial goes on to argue that the experience of Georgia, which has close to a universal pre-K program, shows that universal pre-K is ineffective:
“Careful work by Maria Donovan Fitzpatrick of the Stanford Institute for Economic Policy Research looked at student achievement in Georgia as its pre-K program phased in. While she found some modest gains, she also concluded that the costs outweighed the benefits by a ratio of six to one. Nearly 80% of enrollment is “just a transfer of income from the government to families of four year olds” who would have attended preschool anyway.”
My comment: I have recently re-read Fitzpatrick’s article, which was published in 2008. I have prepared a fuller analysis of what I think the article shows, but I have first shared this analysis with the author to allow a chance to respond before I post this full analysis. In brief, I believe the Wall Street Journal has significantly misinterpreted what the empirical findings of Fitzpatrick’s article show, in several respects:
(1) The six to one ratio for costs to benefits is for a fiscal impact analysis that only counts gains in government revenue (from the increased earnings of former preschool participants) as benefits. Yet a full benefit-cost analysis of any public policy should consider effects on all groups in society, not just the government. If one adds in benefits of preschool for after-tax earnings of former preschool participants, as well as social benefits from reduced crime, benefits could well exceed costs.
(2) Fitzpatrick’s paper presents a wide range of estimated preschool effects using different methodologies, and these estimates vary widely in what they show about the plausible range of effects of Georgia’s preschool program. Depending upon what estimates are used, one can get widely varying benefit-cost ratios for Georgia’s preschool program, which differ in their bottom line conclusion about whether benefits exceed costs.
(3) Fitzpatrick’s estimates include significant effects on both disadvantaged children and non-disadvantaged children.
The Wall Street Journal goes on to look at national trends and evidence on preschool:
“Nationwide today about 1.3 million kids, or 28% of all four-year-olds, attend state-funded pre-K, a leap from 14% in 2002. The empirical case for this expansion—the evidence that universal preschool “works,” as Mr. Obama put it—rests on two academic studies, the Abecedarian and Perry projects, conducted four and five decades ago.”
My comment: The empirical case for “universal” preschool does not rest solely on Abecedarian and Perry. These studies are targeted and small-scale, so they hardly could provide a complete case by themselves.
The empirical case for “large-scale” preschool – which might be either targeted on the disadvantaged, or universal –rests on many other studies that have looked at large-scale preschool programs, including the Chicago Child-Parent Center program, and many state pre-K programs. These studies show that preschool can work on the large scale in the short-run (the state studies) and in the long-run (the CPC study).
The empirical case for more universal programs rests on state studies that have looked at programs with broad eligibility (notably Oklahoma and West Virginia), as well as studies of Tulsa’s program. My study with Gormley and Adelstein of Tulsa’s universal preschool program explicitly compares short-run effects of preschool on middle-class vs. low-income children. We find similar effects on kindergarten entrance test scores, which would be projected to cause similar dollar effects on adult earnings.
The Wall Street Journal goes on to describe these experiments’ effects and costs, compared to Georgia:
“These experiments showed vast returns on investment, the source of Mr. Obama’s claim that every early education dollar generates $7 down the line. Yet Abecedarian and Perry cost between $16,000 to $41,000 per child per year (in current dollars), the higher end comparable to Ivy League tuition. Georgia spends $4,298 per child.”
“The extra money was required because these were very intensive interventions that included home visits, parent counseling, nutrition, health care and other social services. They were micro-enterprises run by the most experienced early education experts and impossible to replicate. Mr. Obama is simply pocketing their results and pretending that this can be extrapolated to the entire population. It can’t even be replicated in Georgia.”
My comment: The implication that we cannot get strong results even with cheaper programs that are focused on preschool services is incorrect. Most of the state and local preschool programs that have shown short-run and long-run benefits cost about $5,000 per year per child. For example, my Tulsa study with Gormley and Adelstein found favorable results for a half-day preschool program for 4-year-olds that cost $5,080 per student in 2012 dollars. Georgia might be under-funded compared to what one would prefer for a high-quality preschool program, but the additional needed funding might be no more than another 20% per student.
The Wall Street Journal then goes on to claim that:
“…What “study after study” really suggest is that government-funded pre-K programs are best when they are targeted at low-income, disadvantaged or minority children—those with the most need. Such a modest, practical reform may lack Mr. Obama’s preferred political grandeur, but the other reason he didn’t propose it is that the government has already been doing it for a half-century. “
“That would be Lyndon Johnson’s Head Start program, birth date 1965. In December of last year, the Health and Human Services Department released the most comprehensive study of Head Start to date, which took years to prepare. The 346-page report followed toddlers who won lotteries to join Head Start in several states and those who didn’t through the third grade. There were no measurable differences between the two groups across 47 outcome measures. In other words, Head Start’s impact is no better than random. “
My comment: First, study after study does NOT show that government-funded-pre-K programs are best when targeted. Most studies focus solely on the disadvantaged, and therefore provide no evidence for or against benefits for the non-disadvantaged. These studies do show strong benefits of high-quality programs for the disadvantaged. Those studies that include the non-disadvantaged, as I have cited above, also show benefits for that group as well.
Returns may be higher for the disadvantaged group, and there may be more evidence of benefits for the disadvantaged, but there may still be net benefits from expanding preschool from disadvantaged groups to the middle-class. As I show in chapter 8 of my book, Investing in Kids, even if we assume significantly lower benefits of preschool for the middle class than for the poor, universal preschool may have higher net economic benefits than targeted preschool. If we want to affect the quality of our entire labor force, then involving middle-class kids in government educational investments is sensible.
Second, I have already extensively commented in a previous blog post on what the recent evidence from Head Start shows, and how it fits in with other previous research. What the Wall Street Journal overlooks is that there are good long-term studies of Head Start, with good comparison groups, that show long-term benefits of Head Start. One such study, by Deming, shows strong long-term benefits of Head Start even though the study also shows considerable fading of the program’s test score effects as students progress in K-12. One possible interpretation of this fading and re-emergence is that it is due to program effects on social skills (“soft skills”) that are more difficult to measure with standardized tests.
The Wall Street Journal again:
“Preschool activists explain away such results by claiming that different programs vary enormously in quality. The White House claims fewer than three in 10 kids are in a “high quality” program. Since we don’t live in Lake Wobegon, well, of course. But it turns out that there are even deep disagreements in the early education literature about how to improve quality, or even how to measure quality in a valid, objective, reliable and fair way.”
My comment: Despite disagreements about how to measure and improve preschool quality, it appears that it is possible for places as diverse as Chicago Public Schools and Tulsa Public Schools to run quality preschool programs that produce large benefits. It is certainly feasible for regular public agencies to run quality programs. And as reviewed in chapter 5 of my book, we do have considerable evidence on what makes for quality in preschool. Quality requires reasonable class sizes, well-trained teachers paid a competitive wage, and a curriculum that emphasizes both academic skills and social skills.
Continuing with the Wall Street Journal:
“Counting Head Start, special education and state-subsidized preschool, 42% of four-year-olds are now enrolled in a government program. Federal, state and local financing for early learning is closing in on $40 billion a year, double what it was a decade ago. But can anyone say that achievement is twice as good—or even as good?”
My comment: Actually, over the last 15 years or so, trends in student achievement at fourth grade, particularly in mathematics, are up. For example, over the period from 1996 to 2011, fourth grade math achievement in the U.S., according to the National Assessment of Educational Progress, has increased by about 0.50 in “effect size” units, which is a gain of about 19 percentiles or about one grade level in performance. Reading scores at 4th grade have gone up about as third as much, with an effect size gain from the mid-1990s to the present of about a 0.15 effect size.
It seems unlikely that all of this trend or even most of this trend has been due to preschool. The 2011 assessment results would reflect students in preschool in 2006. In 2006, 20% of all four-year-olds were in state-funded pre-K programs. The mid-1990s test score results would include students who were in preschool around 1991, when about 12% of all four-year-olds were in state-funded pre-K. Therefore, the increase over this period in four-year-olds in state-funded pre-K is perhaps 8% out of all 4-year-olds. State-funded pre-K might increase achievement by an effect size of 0.35. An increase in achievement by an effect size of 0.35 for an extra 8% of the population would increase overall achievement by an effect size of about 0.03.
Therefore, preschool might explain between 5% and 20% of the test score improvement for 4th-graders since the mid-1990s. Why not more? In large part, because while preschool participation increased by 75% over this time period, the percentage of all 4 year olds in state-financed pre-K still only increased by 8% of all four-year-olds over this time period. The increase in state pre-K is large in terms of program budget increases, but is still modest compared to the size of the entire population.
In sum, there is considerable research evidence that large-scale state-funded pre-K programs can be effective for many students. Although many details remain to be explained, President Obama’s proposal seems to aim at expanding these state pre-K programs, not at implementing Head Start for all. Preschool can not by itself solve all social problems. But if implemented in a high-quality manner on a large-scale, it can make a significant difference in helping develop greater opportunities for many children.