Update on my short videos on early childhood programs: 4th video added, and videos now available on ITunes

Posted on November 5, 2013 by timbartik

With production by Detroit Public TV, and funding from Michigan’s Early Childhood Investment Corporation, I have been working on six videos on early childhood programs. The videos are short (about a minute and a half each). They focus on key issues related to the research on the benefits of  early childhood programs: why should we believe the research evidence for benefits, how early childhood programs have such large long-run effects, and why it makes sense from a perspective of enlightened self-interest to pay higher taxes to pay for pre-K for someone else’s children.

The fourth video is now available. This video discusses why early childhood programs not only pay off for the nation, but also for the state that provides the investment dollars.

In addition, all the videos can now be downloaded for free from ITunes. This should facilitate use at conferences – you don’t have to hope that your internet connection isn’t interrupted by some glitch, because you can download the entire video to your computer.

The remaining two videos in this series should be available soon.

I hope that these videos prove to be useful.  I’d appreciate it if people can let me know if they use the videos, and their assessment of the videos’ utility.

Posted in Early childhood programs, Economic development | 4 Comments

Early childhood education: the economics of early versus later interventions

Posted on October 28, 2013 by timbartik

Nick Kristof had an excellent column on early childhood education in the October 27th edition of the New York Times. In this column, he argues that early childhood education is “the best tool we have to break cycles of poverty”.

In two previous blog posts, from 2011 and 2013, I have tried to quantify how much early childhood education in general, and preschool in particular, could do to reduce income inequality. (The 2011 blog post was commenting on a previous Kristof column, and the 2013 blog post on a Krugman blog post that made similar points about the potential anti-inequality role of universal pre-K.)

What research suggests is that universal pre-K, by itself, might increase future earnings as adults of disadvantaged children by 7% to 15%, depending upon whether we are talking about one school-year of half-day pre-K at age 4, or two school years of full-day pre-K at ages 3 and 4. Moving to high-quality full-day child-care and pre-K for all disadvantaged children from birth to age 5 might add 25% to future adult earnings of disadvantaged children. High-quality parenting programs such as the Nurse Family Partnership might add 3% to the future earnings of disadvantaged children who are part of this program, as well as having a similar magnitude of economic benefits for the disadvantaged first-time mothers who are served by NFP.

The bottom-line is that if we did all of this for disadvantaged children — high-quality parenting programs, high-quality child-care, and high-quality pre-K, at a high intensity and quality level — we would essentially make up for the impact of increased inequality from 1979 to the present on the lowest income quintile. Early childhood education by itself, if pursued at full-scale, can make up for the adverse income trends of the last 34 years, at least in terms of the impact on the earnings of the poor.

However, Kristof alludes to a point in his column, and makes a statement in his blog post accompanying the column, that could be misinterpreted. In his column, Kristof states that “The earliest interventions, and maybe the most important, are home visitation programs like Nurse-Family Partnership.” In his accompanying blog post, he expands on this point to state the following:

“…When we hear “early childhood education” we mostly think of pre-K. In fact, the earlier the intervention, the better. Helping pregnant moms avoid substance abuse is highly cost-effective, and then helping them through home visitation programs like Nurse Family Partnership in the first couple of years of life is crucial as well. By the time you get to age 4, it’s a little late, and children are so far behind that they never catch up.”

This can be interpreted as implying that we should prioritize the earliest possible interventions. I think that conclusion is inconsistent with the research.

Here’s the situation:

(1)    The highest benefit-cost ratios for early childhood education programs, particularly if we are focused on benefits for kids, tend to be for pre-K programs. For example, in my book Investing in Kids, I estimated that the ratio of future earnings benefits to costs for universal pre-K might be about 3.8 to 1. (The future earnings benefits are the present value of future earnings benefits.) The ratio of future earnings benefits to costs for a full-time child care and pre-K program from birth to age 5 is about 3.0 to 1, but only a little more than one-third of this is due to effects on the children, the rest is due to what free high-quality full-time child care does for the earnings prospects of parents.  The ratio of future earnings benefits to costs for the Nurse Family Partnership program is about 2.5 to 1, but half of this is due to benefits for kids, half due to benefits for the moms.

This pattern of relatively high benefit-cost ratios for early childhood programs at ages 3 or 4, versus programs at earlier ages (or later ages), has been previously noted by Arthur Reynolds, Judy Temple, and Suh-Ruu Ou. (See Figure 8.5, in chapter 8 of Childhood Programs and Practices in the First Decade of Life.)

(2)    Why does this pattern occur? After all, kids are more malleable at earlier ages, because neurons are forming at a higher rate.  The reason is that at earlier ages, our interventions must use lower ratios of children to adults, so the interventions are more costly.  Pre-K can get away with staffing ratios of 15 to 20 students per two teachers and still provide high-quality services to kids, assuming the teacher and curriculum is high-quality.  Child care programs prior to age 3 must use smaller child-to-staff ratios to be high quality.  Parenting programs such as NFP provide one-on-one parent visits using skilled nurses. The more individualized services that are required at earlier ages are more costly per child.

Benefits do go up from adding more staff-intensive earlier interventions. However, the benefits do not go up in percentage terms as fast as costs expand. Therefore, the benefit-cost ratios tend to be somewhat lower for the earlier-age interventions.

Pre-K seems to operate in something of a “sweet spot”. Children at ages 3 and 4 are malleable enough that high-quality interventions can make a significant difference. Yet children at these ages are ready to benefit from interventions that can take place in larger groups, which makes interventions less costly per child, or at least less costly per hour of service to the child.

(3)    At the same time, adding on these earlier interventions, even though it tends to lower benefit-cost ratios, still passes a benefit cost test. Even if we already had universal pre-K at ages 3 and 4, adding high-quality child care and parenting programs for disadvantaged families would pass a benefit- cost test.  (The available evidence suggests that such child care and parenting programs, at least in terms of benefits for children, probably are of much more benefit to children from lower-income families that for children from middle-income families. In contrast, universal pre-K tends to have similar dollar benefits for future adult earnings for children from low-income and middle-income families. )

(4)    Furthermore, at noted above, if we really want to deal fully with the earnings gaps of low income families, universal pre-K is not enough. A half-day of pre-K at age 4 raises future earnings by 7%, which is a very large benefit compare to modest costs of perhaps $5,000 per child, but does not come close to making up for growing income inequality. If we really want to deal with poverty, more intervention is necessary per child, which of course is more costly per child.  High-quality child care and parenting programs are interventions that provide valuable add-on benefits.

We need the most intensive early interventions, which begin the earliest, to fully address the challenges of childhood poverty. But policymakers must consider costs as well as benefits.  If budgetary or political constraints prevent fully addressing child poverty, pre-K programs represent a good program area for a start, as such investments can have the highest benefit-cost ratios. However, we shouldn’t stop there, but rather add earlier interventions as budgets and politics allow.

Posted in Distribution of benefits, Early childhood program design issues, Early childhood programs | Comments Off on Early childhood education: the economics of early versus later interventions

Using test scores to evaluate early childhood programs does not imply that they should be used for accountability purposes for individual program centers or teachers

Posted on October 14, 2013 by timbartik

In some of my blog posts and published articles, I have used effects of early childhood programs on early test scores to evaluate programs.  For example, in my Tulsa study with Gormley and Adelstein, we estimated the effects of Tulsa’s pre-K program on kindergarten entrance test scores.  We then used these kindergarten test scores to simulate the likely impact of Tulsa’s pre-K program on adult earnings.  This simulation was based on estimates from Chetty et al. on the likely relationship between kindergarten test scores and adult earnings. (Chetty et al. “How Does Your Kindergarten Classroom Affect Your Earnings?”) With adult earnings effects, and the costs of the program, we then estimated a partial benefit-cost ratio: the ratio of the present value of the earnings benefits of the program to the costs of the program. This benefit-cost ratio is partial because it does not include other benefits such as lower crime, but adult earnings benefits are a major proportion of the likely long-run benefits of pre-K programs for former child participants and for society.

I think the evidence suggests that early test scores are a sufficiently good predictor of adult earnings effects, at least on average, that this is a legitimate procedure. Given that long-run studies of how early childhood programs affect adult outcomes are quite difficult and costly to do, because of  the long-time period involved, and the problems in identifying causal relationships, it is important to find some short-run indicators that give us a reasonably good idea of long-run effects of these programs. This allows us to use many more studies to see how different types of programs will have different benefit-cost ratios. We can use short-run studies to examine the relative effectiveness of different program designs for children from different groups.  Such short-run studies are cheaper to obtain and get results in a timelier manner.  Therefore, such short-run studies are more relevant to real-world policy decisions.

However, I do not think that this necessarily implies that test score effects of early childhood programs should play any substantial role in holding individual program centers or individual teachers accountable for “results”. Why might such an accountability role for test scores be a bad idea, even if test scores are useful in evaluating some overall program design?

The key problem is that once you use test scores of individual program centers or teachers to reward centers or teachers in some way, you distort the behavior of centers and teachers to focus much more attention on maximizing test scores. This would be fine if test score effects of a program, center, or teacher were all that mattered for determining program benefits. But even though test score effects of a program that is NOT being rewarded directly based on test scores are a good predictor of adult earnings, this does not mean that all that matters about the program are its effects on test scores. There is much else in program effects that probably matters. Test score effects may help proxy for these other program effects if the program’s operations are not artificially distorted by accountability incentives, but this is likely to be much less true if centers or teachers are simply “teaching to the test”.

For example, it is commonly believed in pre-K research that much of the long-run benefits for adult earnings and reduced crime of pre-K are due to the program’s effects on “soft skills”: social skills and character skills.  These better soft-skills at kindergarten entrance help students get along better with teachers and classmates in kindergarten, and therefore learn more in both soft skills in kindergarten. And so on through the K-12 system and on into adulthood. In adulthood, these better soft skills are very important in determining success in the labor market.  A person who can get along better with supervisors, co-workers, and customers is likely to retain their job longer, and be promoted to higher-paying jobs.

Why then might it be OK to use effects on kindergarten tests to estimate the effects of a program on adult earnings? After all, these tests mostly measure “hard skills”: various literacy and math skills. Soft skills are hard to measure.

The reason that using test scores is OK for predicting adult earnings is that effects on the hard skills that are measured by these tests are often a reasonable proxy for broader effects of the program on both hard skills and soft skills.  For example, pre-K programs that have more-experienced and better-trained teachers with smaller class sizes and a good curriculum are likely to have stronger effects on both hard skills and soft skills.  We can measure the effects on hard skills, and this is a good indicator of these broader effects.

But effects on test scores may no longer be a good indicator for soft-skill effects and long-run benefits if the behavior of individual centers and teachers is distorted by an undue emphasis on rewarding individual centers and teachers for increasing test scores. This reward for higher test scores may encourage individual centers and teachers to distort what they do to unduly emphasize how many letters or numbers the child knows at kindergarten entrance, rather than the child’s creativity, planning ability, and ability to be both a leader and a team player.

(In theory one could argue that if an entire statewide pre-K program knows it might be evaluated by some outside evaluator periodically based on test scores, this puts pressure on the overall system to overweight tests. But if these test score evaluations are combined with some other evaluation measures of the program, and a conscious decision is made that state program officials should be discouraged from pushing teaching to the test, then this danger is minimal. In contrast, if we set up some accountability system on autopilot with test score measures directly responsible for rewarding individual centers and teachers, there is a much greater danger of distorting behavior.)

How then, can we hold individual centers and teachers accountable for providing high-quality services? Answer: by using that magical thing called human judgment, but hopefully disciplined to some reasonably objective standards, and allowing for multiple sources for such judgments.

For example, a recent article in Science found some evidence that the Classroom Assessment Scoring System (CLASS) gave measures of quality of pre-K classrooms that were strongly correlated with the study’s measure of average test score effects.  These CLASS measures are based on what trained classroom observers see about the quality of teacher-child interactions in the classroom. The CLASS measures, if done with properly trained personnel, seem to be fairly consistent across different observers.

CLASS measures seem likely to be harder to game than test score measures. To improve your CLASS score, you have to improve the quality of teacher-child interactions.  For accountability measures, CLASS measures also have some other advantages compared to test score measures. CLASS measures tell you something about classroom quality now, rather than waiting until the beginning of kindergarten or the end of pre-K to see what the test score effects are. CLASS measures also naturally lead to conversations about how to improve the quality of teacher-child interactions, which seems to be at the core of what quality means in pre-K.

Using test score effects for accountability in pre-K is certainly worth experimenting with. Maybe it can work if the test score measures are not over-emphasized, and if other quality measures, such as CLASS, are also used. But I think there are clearly dangers in over-emphasizing test scores in accountability measures.  I don’t think we know enough yet for the federal government, state governments, or even local governments, to mandate that test scores should be used for some percentage of an accountability metric for individual pre-K centers or teachers, unless such a system is part of an experiment or demonstration project.  I suspect that as we learn more about the best accountability system for pre-K quality for individual centers or teachers, that the accountability systems that work the best will emphasize measures such as CLASS much more than test scores.

Posted in Early childhood program design issues | 3 Comments

New short minute and a half videos (with animation!) on why preschool makes sense

Posted on October 9, 2013 by timbartik

I have been working with Detroit Public Television and Michigan’s Early Childhood Investment Corporation on preparing short videos about early childhood programs, including preschool. The project was initiated and financed by ECIC; I prepared the scripts; Detroit Public TV did an excellent job of creating the animation and professionally editing each video.

Three videos are available for viewing now. They are at an alternative website about some of my research, at http://www.investinginkids.info/ .

Each video is about a minute-and-a-half long. Each video considers a different topic.  Here are the short summaries of each video and a specific link to each video:

Video 1: “Why Preschool Matters”. “How is it even possible that just one year of preschool can produce such large, long-run benefits? Dr. Bartik explains how just one year of preschool lays the foundation for a child’s lifelong learning and directly impacts their productivity and earnings as an adult.”

Video 2: “Research on Preschool”.  “Why should I believe that early childhood programs work? There have been many studies researching the impact of early childhood programs comparing those who attend and those who don’t. Dr. Bartik takes a close look at the two groups and explains the surprisingly large differences between the two groups.”

Video 3: “Your Taxes and Preschool”.  “Why should I pay higher taxes for other people’s children to attend preschool? No one wants to pay higher taxes, but Dr. Bartik explains how everyone in today’s society is interdependent. Watch to find out why children attending preschool will benefit the local economy and how this has a direct and prosperous effect on you as well.”

These videos are intended to be disseminated as widely as possible. Feel free to send links to videos whenever you think it might be helpful. Please also feel free to use these videos at conferences. If possible, if you could let me know that you’re using the videos at a conference, that would be helpful.  You can do so at the comments at this blog post, or by emailing me at bartik@upjohn.org.

A few more videos are still in production, and I will announce these new videos as they become available.

I hope these videos are helpful in providing short, clear, and entertaining answers to some of the questions about the benefits of preschool.

Posted in Early childhood programs | 2 Comments

New evidence suggests universal pre-K increases overall pre-K enrollment for all income groups, but provides mixed evidence on test score effects

Posted on October 2, 2013 by timbartik

A new paper by Elizabeth Cascio and Diane Whitmore Schanzenbach provides important new evidence on the effects of universal pre-K. This paper was presented in late September at the Brookings Panel on Economic Activity.

Cascio and Schanzenbach examine the effects of universal pre-K in Oklahoma and Georgia. The focus is on effects on overall pre-K enrollment (that is, total enrollment in both private and public pre-K programs), and on effects on 8th grade test scores.

A possible concern about universal pre-K is that its availability may be mostly or entirely offset by reduced enrollment in private pre-K. If this “crowd-out” of private pre-K occurs, then this may reduce the net overall effect of universal pre-K on participants’ skills. For a participant who enrolls in public pre-K, but otherwise would enroll in private pre-K, the net effect of the universal pre-K program on skills may be reduced. Perhaps the private pre-K would have increased skills almost as much as the universal pre-K program, or by even more.

Although Cascio and Schanzenbach find some crowd-out, it is not close to being 100%. Universal pre-K significantly increases overall pre-K enrollment.  This enrollment increase due to universal pre-K occurs for both disadvantaged students, and advantaged students.  The crowd-out effect is somewhat larger for advantaged families, but there is still a sizable net increase in enrollment for this group.

Cascio and Schanzenbach also find that universal pre-K increases 8th grade test scores for disadvantaged students, compared to what would have occurred in states without universal pre-K. For more advantaged students, they do not find statistically significant effects of universal pre-K on 8th grade test scores.

Some media reports have spun these results as showing that universal pre-K does not provide long-term benefits to middle-income children.  A more balanced interpretation is that the Cascio and Schanzenbach results do not provide strong evidence either for or against many of the most important potential long-term effects of pre-K programs for the middle-class.  A particularly important long-term economic effect of pre-K is effects on former participants’ adult earnings. 8th grade test score effects are unlikely to be strongly predictive of whether or not there will be future adult earnings effects.

Several good studies suggest that 8th grade test score effects are not a good indicator of whether an early intervention affects adult earnings. Chetty et al. find that kindergarten “class quality” has effects on test scores that fade, but that effects of kindergarten class quality on adult earnings are still significant.  (The Chetty et al. paper is “How Does Your Kindergarten Classroom Affect Your Earnings?”, and can be found on this page.) Chetty et al’s results find that effects of kindergarten class quality on kindergarten test scores can be used to predict reasonably accurately the eventual effect of kindergarten class quality on adult earnings.  But kindergarten class quality’s effects on 8th grade test scores are so small that they predict effects on adult earnings that are less than one-sixth of the adult earnings effects that we eventually observe.

Deming finds that Head Start has sizable effects on adult earnings, even though effects on test scores fade to statistical and substantive insignificance by 8th grade. Head Start’s initial test score effects decrease by over 60% by ages 11-14, and are no longer statistically significant. Despite this fading of test score effects, Deming finds significant Head Start effects on adult outcomes that are sufficient to predict that Head Start increases adult wages by 11%.

What might explain this fading and re-emergence of the effects of early intervention? One possible interpretation is that many early interventions significantly increase both “hard skills” (academic skills) and “soft skills” (social and character skills) in the short-run. Some of the hard skill effects fade over time, but the soft skill effects continue, and lead to better adult outcomes.  Soft skills are hard to directly measure, and so most school test scores really measure hard skills. The initial hard skill effects on early test scores of early interventions may indicate an increase in both hard skills and soft skills due to a well-run program, and may be a better predictor of long-run adult earnings effects than the faded hard skill effects that are observable in later grades.

In sum, I don’t think that Cascio and Schanzenbach’s new research is inconsistent with the previous research literature on universal pre-K. This previous research has found that universal pre-K has strong effects on early test scores for both low-income and middle-class students.  My paper with Gormley and Adelstein on Tulsa’s universal pre-K program found that this program had similar early test score effects for both low-income and middle-class students. We used these early test score effects, and the links developed by Chetty et al. between early test scores and adult earnings, to project that Tulsa’s program would have similar dollar effects on increasing adult earnings for both low-income groups and middle-class groups.  Weiland and Yoshikawa’s recent study of Boston’s universal pre-K program also finds similar effects on early test scores for both low-income and middle-class children. In a recent blog post, I used their results to project that Boston’s program is likely to have similar dollar effects on adult earnings for both low-income groups and the middle-class.  Cascio and Schanzenbach’s new research does not provide strong evidence for or against these projections that universal pre-K will have similar effects on adult earnings for both the poor and the middle class.

Cascio and Schanzenbach’s study points out that one good rationale for universal pre-K is the possibility of significant peer effects. If having more middle-class children in a pre-K classroom tends to provide positive peer effects that enhance the learning of low-income children, then part of having quality pre-K for the disadvantaged is having income-integrated pre-K classrooms.  Income-integrated classrooms are much easier to implement under a universal pre-K system than under an income-targeted pre-K system.

The weight of the research evidence continues to suggest that universal pre-K has broad benefits for many income groups.  This research evidence, along with the potential broader political support for universal pre-K, are good reasons to support making expanded pre-K accessible to as many families as possible.

Posted in Distribution of benefits, Early childhood program design issues, Early childhood programs, Timing of benefits | Comments Off on New evidence suggests universal pre-K increases overall pre-K enrollment for all income groups, but provides mixed evidence on test score effects

How much can universal pre-K do to reduce income inequality?

Posted on September 13, 2013 by timbartik

In his September 13, 2013 column in the New York Times, Nobel-prize-winning economist Paul Krugman discussed the growing problem of increased American inequality.

As one solution, Professor Krugman mentioned the idea of “universal prekindergarten education”, in order “to expand opportunity for the children of the less fortunate”, which would “do at least a bit to level the playing field”.

A natural question is:

How much can universal pre-K do to reduce income inequality?

The brief answer: A large amount for the lowest income quintile and middle income quintile, but it is still not a complete solution.

A more complete solution for the lowest income quintile would add further redistributive efforts which could include developmental child care programs such as Educare or parenting programs such as the Nurse Family Partnership.

A more complete solution for the middle income quintile would consider other measures that provide public services that are both productive and at least modestly redistributive, such as programs to improve public education and increase access to higher education. We should also consider how to run the economy at a lower unemployment rate to put upward pressure on wages.

The best recent research on pre-K suggests that universal pre-K would cause similar percentile increases in test scores for different income groups. This research comes from both my research with Gormley and Adelstein on Tulsa’s pre-K program, and Weiland and Yoshikawa’s research on Boston’s pre-K program.

Based on the research of Chetty and his colleagues, we would expect these similar percentile increases in test scores to result in similar DOLLAR increases in adult earnings for all income groups. But the key point is that a similar dollar increase in adult earnings for all income groups is still a huge redistribution of income.

In my 2011 book, Investing in Kids, I projected that if universal pre-K has equal dollar effects on all income groups, then even one school year of half-day pre-K would boost the income of the lowest income quintile by 6%, and the income of the middle income quintile by 2%. (See Table 8.4)

In our analysis of Tulsa, we similarly found that one year of half-day pre-K would boost the future earnings of children eligible for a free lunch subsidy by 7%. Children whose family income was too high to be eligible for any lunch subsidy would have future earnings increased by 3%.

A full-day pre-K program would do more. In Tulsa, we found such a program would boost future earnings of the free lunch eligible group by 10%, and for the higher income group by between 5 and 6%.

Adding a second year of pre-K would do even more. The available evidence (see chapter 5 of my 2011 book) suggests that a second year of pre-K might multiply the one-year earnings effect by 1.5.  So we might find that two years of full-day pre-K would boost the lowest income quintile’s earnings by up to 15%, and the middle income quintile by 8 or 9%.

How much of a dent does this make in the increased income inequality in the U.S.? In a previous post, I calculated that over the last 30 years, increased income inequality has reduced the income of the lowest income quintile by 40% compared to what it would have been if this income group had equally shared in the benefits of overall U.S. economic growth. For the middle-income quintile, increased income inequality has meant 30% lower income than if this group had equally shared in income growth.

So, universal pre-K, by itself, might fill a considerable portion of the gap for the lowest income quintile and the middle income quintile. How much depends upon whether we are talking about a one-year half-day program vs. two years of full-day pre-K.

The price tag? Universal half-day pre-K for one year has a national price tag of around $15 billion, or $50 per capita. Expanding to a full-day or to two years could blow up this price tag by up to a factor of four. In terms of a U.S. economy with over $16 trillion in annual output, even a price tag of $60 billion is not huge.

For the lowest income quintile, programs such as Educare would do even more. This full-time developmental child care from ages zero to five might boost earnings by 25% or more. The available evidence suggests such a program would only benefit children from low-income families, not children from middle-income families. The price tag for this program is $40 billion, but the program is more politically difficult because it only benefits the lowest-income group.

It is perhaps unfair to expect early childhood education to completely “solve” a huge economic and social problem such as massive increases in income inequality. What is perhaps surprising is how much early childhood education might do on its own to address this issue. Early childhood education is not a complete solution to income inequality, but it would be a good start.

Posted in Uncategorized | 1 Comment

The potential of parenting programs, as well as pre-K programs

Posted on September 11, 2013 by timbartik

A recent report by the Center on Children & Families of the Brookings Institution has some useful information on parenting in the United States, and some useful research results for the HIPPY parenting program. The report is by Richard V. Reeves and Kimberly Howard, who are respectively Policy Director and Research Assistant at the Center.

Unfortunately, the report is also marred by a few paragraphs on pages 1 and 2, and page 16, that pit parenting programs against pre-K programs.  These few paragraphs have already drawn attention from Susan Ochshorn of ECE PolicyWorks. These paragraphs are not well backed by the available research. However, before getting to those paragraphs, let’s focus on the main content of the report, on what parenting programs can do.

The report documents that there are a great many issues that might be raised with parenting in the U.S. These issues are based on home observation of the cognitive and emotional environment provided by parents. Perhaps the key finding is that there is a much bigger gap, all else equal, between the quality of the home parenting environment for low-income families versus middle-income families than between middle-income families and upper-income families.

This should not be surprising. Low-income families face many stressors such as inadequate wages, irregular employment, and unstable housing situations, all of which create emotional and time pressures on parents and in turn stress their children. As I’ve mentioned before, research by Greg Duncan and his colleagues provides good evidence that income increases, especially in the early childhood years, have large effects on improving adult outcomes for children.

As the Brookings report documents, research suggests that parenting problems will contribute to higher high school dropout rates and other problems with adult outcomes. If we could somehow wave a magic wand and improve parenting, this would help improve these outcomes, although they appropriately add that “better parenting is very far, on its own, from being a magic cure”.  (p. 10).

The Brookings report also documents that both cognitive and emotional aspects of parenting matter to later outcomes. This is not a surprise to researchers, or indeed, to any parent or to the average person reflecting on his or her own life experience.

The policy issue is what, if anything, can be done about problems with parenting, in the absence of a magic wand. The Brookings Report focuses on home visiting programs, under which the program provides a number of home visits during some phase of early childhood to the parent and perhaps child, with an attempt to provide support, help, advice, and modeling of some better parenting practices.

The Brookings report states that “most home visiting programs have shown disappointing results”. (I might have said “many” because it is unclear to me how one counts successes and failures, but it is certainly the case that not all home visiting programs work.) However, they argue that programs can work if they are intensive and have well trained and supervised home visitors.

In particular, the Brookings report cites the well-known research evidence for the success of the Nurse Family Partnership.  In addition, they cite the evidence compiled by Mathematica Policy Research that sought to identify successful home visiting programs that would be eligible for federal funding under the Affordable Care Act.

The Brookings report goes on to focus on the potential effects on adult outcomes of the Home Instruction for Parents of Preschool Youngsters (HIPPY) program, which was identified by Mathematica as a successful program based on research.  The Brookings report describes HIPPY as working with parents of children ages 3 to 5 by biweekly home visits for 30 weeks per year and biweekly group meetings, along with the provision of some educational books and toys.

The Brookings report uses short-term results from HIPPY on test score in first grade to estimate effects on high school graduation. (This simulation is based on the Brookings Social Genome Model.) The estimate is that the program might increase high school graduation rates from 75.4% to 77.5%, or 2.1%.

How does this translate into benefits and costs? According to Brookings, HIPPY costs about $3,500 per child. They cite research that says a high school graduate earns about $260,000 more than a dropout. My own research in chapter 12 of my book, if recalculated on a national scale, implies that the present value as of age 4 of a high school graduate’s earnings are about $318,000 more than a dropout. (Wonk note: This recalculates the $175,000 of Table 12.1 in chapter 12 without any out-migration.)

If we multiply $318,000 by 0.021, the change in the probability of getting a high school diploma, we get an earnings gain of $6,678 per the “average” HIPPY participant. The ratio of the present value of the earnings gain to the cost of the program is 1.91. (Alternatively, if we ran this program for 1000 children, we would have costs of $3.5 million. 21 extra children would graduate from high school, for a gain of $6,678,000 (=$318,000 times 21) in present value of earnings.)

All of this is good and valuable research. Parenting is an issue for some low-income families, and some programs can help, with good benefit-cost ratios.

However, the report for some reason decides to have a few paragraphs at the beginning and end of the report that contrasts “parenting programs” with “early childhood education programs” such as pre-K.  Personally, it seems to me that parenting programs are a form of early childhood education, but they choose to make a big contrast as to whether the program primarily interacts with the parent or the child. In the real world of quality pre-K programs and quality parenting programs, both sets of programs include significant interactions with both child and parent.

The report includes the rhetoric that early childhood education programs seek “to make [parents] less relevant” (p. 1) and “The goal is, in effect, to detach the opportunities of the child from the abilities of the parents.” (p.2).

The same rhetoric could of course of applied to any educational intervention, including all of K-12 education. (For that matter, it could be applied to ANY program that benefits children in any direct way – recreation programs, etc.) But this is wrong-headed in that quality pre-K programs, for example, aim at complementing whatever it is that parents do to improve their child’s development. Quality pre-K programs provide educational services such as socialization, and learning in a group context, that would be quite difficult for even the best parents to provide on their own.   Quality pre-K programs also seek to inform parents of what the program is doing and to involve parents in the child’s progress.

It is very hard to see how a voluntary pre-K program that operates for half-a-day at age 4 for one school year, or even for two school years full-day at ages 3 and 4, is somehow doing anything to hurt parents. It is helping children from all families. In that sense, we can say that the program is making the handicaps of income and poor parenting less relevant. But for a good parent, or even any parent who intends for the good for their child, which I think is the overwhelming majority of all parents, what pre-K programs are doing to help child development does not diminish the parent as a human being. Rather, it helps the parent achieve what she or he wants, which is a better future for the child.

They combine this with rhetoric on page 2 that goes after Head Start as representing all pre-K programs: “A particular disappointment is that Head Start, a flagship early childhood program, appears to be having no measurable impact on academic performance through third grade”.

I’ve addressed Head Start in previous posts. But briefly: (1) there is considerable rigorous research showing long-term effects of Head Start;  (2) even after test scores effects fade, there is research by Deming that suggests large Head Start effects on adult wages;  (3) Head Start is far from the only pre-K program, in that we have state and local pre-K programs of a wide variety.

As I’ve outlined, based on kindergarten entry test score effects, my paper  with Gormley and Adelstein on Tulsa pre-K predicts adult earnings effects whose present value is from 2.82 to 4.08 times the cost of the program. The highest returns are for half-day pre-K, which costs $4403 per child, and increases the present value of earnings for the lowest income children by 4.08 times this cost. Similar high returns are implied by a recent study of Boston pre-K.

If one wants more direct measures of earnings gains than these projections based on early test scores, as done both in my own work and in the Brookings report, this can be found in the various benefit-cost analyses of the Chicago Child-Parent Center program, which was a half-day pre-K program for age 4 for about half the sample, and a half-day pre-K program for ages 3 and 4 for the other half of the sample. We now have follow-up on adult outcomes through ages 26 and 28, not just test score effects at first grade or kindergarten.   The overall benefit-cost analysis as of age 26 for CPC suggests that it increases the average present value of the child’s future earnings by $28,844, at a cost that averages $8, 512 per child. The implied ratio of earnings effects to costs is 3.39. The benefit-cost ratios appear to be about 30% greater for children who only attend the program for one year, due to costs declining by a higher percentage than benefits.

In its next to last paragraph, the Brookings report makes the following arguments:

“Currently, the U.S. spends significantly more on pre-K education than on parenting  programs. In the last 5 years (2009-2013), the federal government has allocated $37.5 billion to Head Start, 25 times the $1.5 billion that it has allocated to evidence-based home visiting programs over the next five years. The Obama administration is proposing an increase in investments on both. But the relative weight of policy remains strongly on the side of supplementing parenting, rather than improving parenting. The analysis presented in this paper suggests that parenting may be worthy of a greater share of public investment. There is strong evidence that parenting influences child outcomes, and some evidence that good programs can improve parenting.”

So, apparently the policy choice is either: fully fund Head Start, or divert some of its funds for parenting programs that show promising results.

Why is the choice phrased in this way? Let’s do some division. If we divide $37.5 billion by 5, we get $7.5 billion per year for Head Start. This is hardly an exorbitant sum of money in the context of the entire U.S. economy. U.S. GDP is currently over $16.5 trillion annually, so the Head Start spending is 0.05% of GDP, or 5 pennies in $100.  Total federal spending is over $3.5 trillion annually, so the Head Start spending is 0.2% of total federal spending, or 2 pennies out of $10.  

In other words, why don’t we expand both research-proven pre-K programs AND expand research-proven parenting programs? For that matter, why don’t we expand research-proven child care programs with educational components, such as the Educare program? Where should the money come from? Quite simply, by either increased taxes, or by reallocating priorities in government budgets.

In my book, Investing in Kids, I estimated that a program that did all this might annually spend an extra $14 billion for half-day universal pre-K, $40 billion for an Educare style program, and about $4 billion for proven parenting programs such as the Nurse Family Partnership.  The total price tag is $58 billion per year.  

This cost amounts to about $200 per capita. This is a cost that could in fact be absorbed by state governments without any federal action.  A sizable piece of it could also be paid for locally if state governments were willing to authorize local policy initiatives.

There is significant research evidence that a wide variety of early childhood educational interventions can work, including high-quality parenting programs, high-quality developmental child care, and high-quality pre-K.  These programs complement each other, as well as complementing what parents do on their own. We need to expand all these programs as warranted by the research evidence on benefits and costs, rather than fighting over a fixed government budget for programs affecting children. 

Posted in Early childhood program design issues, Early childhood programs | 3 Comments

The benefits of early childhood programs depend upon design

Posted on September 2, 2013 by timbartik

Melinda Wenner Moyer wrote a recent column in Slate on how U.S. day care affects kids. Her column provides a useful summary of the day care research. Day care has mixed effects relative to parental care: more positive if the day care is higher quality, the child is older, the family is lower income, and the focus is on cognitive rather than behavioral skills; more negative if the converse holds.

How can these findings be reconciled with the finding, also backed by research, that investing in high-quality early childhood programs has large benefits, not only for the child’s future, but for the local and national economies?

The key point is that the benefits of investing in early childhood programs depend crucially on the design of the programs that are being invested in.

For high-quality pre-K programs at ages 3 and 4, there is strong evidence that pre-K pays off in a better future for former child participants who come from families with a wide variety of income levels. Furthermore, quality pre-K seems to be deliverable at reasonable cost by state governments and local school districts. Quality pre-K requires significant but not extraordinary costs per child, and a good curriculum and good teachers, but not necessarily genius personnel. Such quality pre-K develops both “hard skills” (academic skills) and “soft skills” (behavior, character).

High-quality child care at ages less than 3 can also yield positive effects on child development. This was most rigorously demonstrated in the Abecedarian experiment, which provided high-quality child care and preschool from birth to age 5. (Abecedarian was quite similar in design to the current Educare program, which is promoted by the Buffett Early Childhood Fund and the Ounce of Prevention Fund.)

However, high-quality child care at ages less than 3 probably provides the greatest child development benefits to children from low-income families. The Abecedarian experiment was restricted to such families. The Infant Health and Development Program (IHDP) provided services similar to the Abecedarian program from ages 1 to 3. Research suggests that IHDP services provide much larger benefits for children from low-income families.

Of course, child care is also intended to provide parents with the ability to work or attend school. Such parental benefits are also important to both parents and the broader society and economy. Parental benefits probably occur for families from a wide variety of income groups.  Therefore, it is not obvious that expanded support for high-quality child care should be limited to low-income families.

Parenting programs also seem to provide greater benefits for children from low-income families. The Nurse Family Partnership has some of the most rigorous evidence for the benefits of home visiting services that seek to improve parenting. Research on NFP suggests its benefits are greater for children from more disadvantaged families.

Finally, the greatest concerns about the impacts of early child care are for infants of less than one year of age or less than six months of age.  If this care is of low quality, and the family is more advantaged, there is some indication that such low-quality care may contribute to some early behavioral issues.

What is going on here? The key is what early childhood programs are providing relative to what families are able to provide on their own. Quality pre-K provides services that are difficult for most families to substitute for with their own resources. Such services including learning social skills and learning cognitive skills in a group setting. In contrast, low-quality child care for infants provides interactions with children that many more advantaged families may be able to provide just as well or better on their own.  (Of course, these advantaged parents may not be able to provide such services to their children and still keep the advantages of working or going to school.)

I think the policy implications of this analysis are straightforward. Public support for pre-K programs with broad access seems sensible. This should be coupled with subsidies to help support higher-quality child care services, particularly for low-income families.  As Jonathan Cohn of The New Republic has highlighted, there are serious problems with the quality of much of U.S. child care. Higher-quality child care will help child development in low-income families, while helping parents in all families. In addition, we should expand research-proven parenting programs such as NFP for low-income families. In summary: universality for pre-K, soft targeting for child care, hard targeting for parenting programs; quality for all programs.

As Magnuson and Waldfogel have recently argued, we should also consider income support to parents with young children as part of early childhood policy. Research by Duncan and others has shown that adult outcomes for children are particularly sensitive to their parents’ income when the child was less than age 5.  Public policies that can increase jobs and earnings, or provide income supplements, to low-income families with young children, can have high benefit-cost ratios because of their benefits for child development and the child’s future.

The details of early childhood policy matter.  Universality is most strongly justified for pre-K programs. Parenting programs should probably be targeted at low-income families. Child care programs are likely to have the most child development benefits for low-income families, although all families may reap benefits for parents. In achieving the benefits of all these types of early childhood programs, quality is the key.  Our support should be for expanding quality pre-K programs and quality child care programs, not for simply expanding any early childhood program. And helping children may sometimes be done best by policies that directly help the incomes of their parents rather than by providing services directly to the children.

Posted in Distribution of benefits, Early childhood program design issues, Early childhood programs | 3 Comments

Pre-K benefits the middle-class as well as the poor in Boston

Posted on August 25, 2013 by timbartik

A recent study of Boston’s universal pre-K program provides additional evidence that pre-K programs have benefits greater than costs for children from middle-class families. This supplements the more extensive evidence that pre-K has high benefit-cost ratios for children from low-income families. This Boston study is by Christina Weiland and Hirokazu Yoshikawa.

Previous research on Tulsa’s universal pre-K program, by me and Bill Gormley and Shirley Adelstein, showed large benefits for both low-income and middle-class families. Test score gains due to pre-K at kindergarten entrance for children ineligible for a free or reduced price lunch were only 10-11% less than test score gains for children whose family income was low enough to qualify for lunch subsidies.

The Boston study estimates similar results. As reported in Weiland and Yoshikawa’s published paper, for most student outcomes at kindergarten entrance, the estimated effects of pre-K did not differ significantly with student eligibility for lunch subsidies.

Christina Weiland graciously responded to an email and sent me an unpublished appendix with subgroup results. From Chetty et al.’s research, we have estimates of the relationship between cognitive outcomes in kindergarten, measured as percentile test score gains, and dollar gains in adult earnings. I used the appendix provided by Weiland to compute the percentile test score gains for children eligible for a subsidized lunch, versus higher-income children who were ineligible for a subsidized lunch. (These are test score effects on four tests, which included vocabulary, letter-word identification, and two math tests.)

I compute, based on Weiland and Yoshikawa’s estimates, that Boston’s pre-K program’s effects on cognitive skills for middle-class children are 29% less than effects for lower-income children.  Percentile gains for children eligible for a subsidized lunch average 21 percentiles across these four tests.  For children whose family’s income was too high to quality for lunch subsidies, the estimated test score gain is lower but still sizable at 15 percentiles.

(Note on calculations: I used the appendix information on how test scores varied with free and reduced price lunch status, pre-K participation, and pre-K participation interacted with free and reduced price lunch status, as well as published information on the standard deviation of test scores at kindergarten entrance.  To compute percentile gains at kindergarten entrance by group, I assumed that the test score distributions were normal, and test score distributions were calculated with the overall average at kindergarten entrance set at the 50th percentile. Calculations were done separately for each of the four tests and then averaged. The average effect size of the higher-income group is 34% below that of the lower-income group. However, because the lower-income group tended to be further away from the entrance mean than the higher-income group, the percentile gain differentials are somewhat less.)

The Boston test score gains for pre-K for middle-class sufficient are large enough to easily pass a benefit-cost test. Boston’s pre-K program appears to be relatively expensive, at perhaps $15,000 to $17,500 per student. (Boston’s program is a full-day program, pre-K teachers are paid at the same average salaries as other Boston Public Schools’ teachers, Boston Public School teacher salaries average over $82,000, and Boston’s pre-K classrooms seem to average a class size of 12.3 students.) But the estimated future adult earnings gains based on percentile test score gains at kindergarten entrance are also large. Based on the percentile gains in test scores from Boston pre-K, the ratio of the present value of the future increase in adult earnings, to the cost of the program, are estimated to be $3.22 for children eligible for a subsidized lunch, versus $2.30 for middle-class children.

(Note on calculations: I adjusted estimated earnings effects from Chetty et al., as used in Bartik/Gormley/Adelstein, for the higher Boston vs. Tennessee prices using the price indices in Aten, as well as adjusting to 2012 prices. This resulted in an estimate that each 1 percentile increase in kindergarten test scores will increase the present value of adult earnings by $2,657 in 2012 Boston prices. I multiplied this by the percentile gains computed above to get a total present value of adult earnings gain from pre-K. For these calculations, I assumed the average pre-K cost per student in Boston was $17,387. This is based on adjusting the numbers for pre-K costs for full-day pre-K that pays public school salaries, in the “Meaningful Investments in Pre-K” study by the Institute for Women’s Policy Research, for Boston’s higher teacher salaries, higher prices, and lower class size. Class size estimates are based on the report in Weiland and Yoshikawa that there were 2938 eligible children in 238 classrooms, although this includes both pre-K and kindergarten classrooms.)

These Boston benefit-cost results are similar to the estimates by Bartik/Gormley/Adelstein for Tulsa. The Tulsa results for full-day pre-K estimated a benefit-cost ratio for future earnings benefits relative to pre-K costs of 3.09 for children eligible for a free lunch subsidy, versus 2.82 for children whose family income was too high to quality for a lunch subsidy.  Pre-K is more expensive in Boston than in Tulsa, but the higher Boston earnings mean the earnings benefits are greater as well.

The Boston and Tulsa studies both suggest that pre-K has broad benefits for many income groups. Pre-K provides services in terms of social skills, as well as cognitive skills, that are hard for many families to develop for their children as well on their own.

For both substantive and political reasons, pre-K programs that provide broad access to many groups make sense. Not only is this politically more attractive, but universal access pre-K programs will provide higher overall benefits for the local and national economy. If we want to truly upgrade the overall quality of the local labor force, we need pre-K programs that provide high-quality services to a high proportion of all children.

Posted in Distribution of benefits, Early childhood program design issues, Early childhood programs | Comments Off on Pre-K benefits the middle-class as well as the poor in Boston

Promising results for relatively time-intensive parenting program

Posted on August 21, 2013 by timbartik

A recent study of a parenting program in Jamaica for highly disadvantaged children has found strong evidence for large effects on adult earnings. The study was written by a large group of researchers, including Paul Gertler, James Heckman, Rodrigo Pinto, Arianna Zanolini, Christel Vermeersch, Susan Walker, Susan Chang-Lopez, and Sally Grantham-McGregor.

Obviously the name that immediately draws attention is Nobel –prize-winning economist James Heckman, although the other researchers are also well regarded.  But leaving aside research reputations, the study is well done and gets strong results. I think the study should lead to some new programs in the U.S. to see if the results can be replicated in a variety of settings.

The program targeted “stunted” children in Jamaica who on average were one and a half years old.  Stunted children were those significantly below normal height for their age.  This stunting is used as an indicator of malnutrition, which we know from other research will contribute to poor cognitive development. The program provided these children with services for 2 years.  The services provided were nutritional supplementation, and “psychosocial stimulation”.  This study and previous studies have not found strong lasting effects of the nutritional supplementation, but have found large and lasting effects of the psychosocial stimulation.

The “psychosocial simulation” consisted of weekly one-hour play visits with the child and the mother in the home by a “trained” (16 weeks of training) paraprofessional health aide.  The health aides showed mothers ways to provide educational enrichment for their child through playing educational games with the child, talking to the child, and helping name objects in the child’s environment. The mother was encouraged to continue these activities throughout the week.  New educational toys were provided on each visit and swapped for toys provided on previous visits.

The program evaluation is based on random assignment. There is relatively modest attrition in the original sample, and the attrition that does occur seems to be random, so the treatment and control groups seem to be truly comparable even after the attrition.  The sample size is small: 53 children in the treatment group and 52 children in the control group could be located for a follow-up at age 22. But the small sample size makes it more surprising that the program could find statistically significant effects. This statistical significance only is possible because the program has such large effects on adult outcomes.

The estimate is that this the two years of weekly home visiting increased adult earnings in the program group by at least 38%. This is a very large effect. It is large enough that the program group largely “catches” up to matched “non-stunted” individuals in adult earnings.  This is remarkable given that the stunted children started out with a huge malnutrition disadvantage.

Some clues to the causes of this earnings effect can be found in the pattern of other outcomes.  The program significantly increased the measured quality of parenting during the two years in which services were provided, but not at later ages.  Therefore, any long-term effects must reflect some special effects of improving parenting and child development during the child’s early years. The program significantly increased former child participants’ long-term educational attainment and cognitive skills, and also increased former child participants’ self-esteem while reducing anxiety and depression.  The program group did NOT catch up with the non-stunted group in cognitive skills, but did catch up with the non-stunted group in educational attainment and various psychological strengths. This suggests that the earnings effect is due at least as much to improving “soft skills” as “hard skills”. Soft skills are at least as important as hard skills in improving educational attainment and adult earnings.

One aspect of this program that is striking is its time-intensity.  The “psychosocial stimulation” involved 100 hours of visits over 2 years. In contrast, the home visiting program in the U.S. that is most similar to the Jamaica program, the Parent-Child Home Program, involves perhaps 50 hours of visits over 2 years. The well-known and well-evaluated Nurse Family Partnership program also involves perhaps 50 hours of home visits over a 2.5 year period (p. 98 in my Investing in Kids book provides estimates and sources). So the program seems to involve perhaps twice the time-intensity of some home visiting programs in the U.S.

In the U.S., home visiting programs have frequently shown evidence of success. The Nurse Family Partnership program has shown the strongest long-term evidence of success, but there are also favorable evaluations for other programs. However, it must be admitted that the evidence for home visiting programs in the U.S. is mixed. Not all studies find success for all programs (see pp. 11-12 of Executive Summary of October 2012 Mathematica Policy Research report for U.S. Department of Health and Human Services on home visiting program evidence).  A natural question is whether greater time intensity of home visiting programs in the U.S. would lead to more consistent statistical evidence for long-term effects.

Another natural question is whether the Jamaica results can be validly extrapolated to the U.S. context. This question deserves careful experimentation and demonstration projects.

What is true is that a U.S. home visiting program on the Jamaica model could pass a benefit-cost test even if it only had a fraction of the earnings effects of the Jamaica program. In the U.S., 100 hours of home visiting would probably have less than a $10,000 two-year cost per parent and child. (Per-year costs for the Parent-Child Home program are around $2,500—see page 9 of study by Dickens and Baschnagel — and there might be economies of scale from doubling home visiting hours.) The average present value of future adult earnings for a child eligible for a free price lunch is at least $250,000. (This takes adult earnings over the child’s entire adult working career, and discounts them back to age 4.  These calculations were done for my paper with Gormley and Adelstein on preschool in Tulsa. The subsidized lunch status of the child is commonly used in education research as a readily available proxy for the child being from a low income family.) With a cost of $10,000, we only need a 4% earnings effect in the U.S. for such a program to have earnings benefits that exceed costs. ($10,000=4% of $250,000). Therefore, a U.S. version of this program would only need to achieve about one-tenth of the effectiveness of the Jamaica program to pass a benefit-cost test.   If the program had other benefits, such as reducing crime, the program could pass a benefit-cost test even with a smaller earnings effect.

One general lesson here is that if early childhood programs have lasting effects on adult outcomes such as earnings, they don’t have to have very large percentage effects to pay off even if the programs seem to be initially costly. Affecting an entire life course of a person has such large potential benefits that even costly interventions can make sense.

We would benefit from some additional demonstration programs and experiments with home visiting programs that might moderately ratchet up the time intensity of these programs.  Two straight years of regular weekly visits might have more of an impact – or at least more of a statistically detectable impact – than less intensive interventions.

There is broad interest from across the political spectrum in programs that might help improve parenting in highly disadvantaged families.  But to improve parenting probably requires programs that are both carefully designed and more intense than some people imagine.  Parenting programs are not a cheap alternative to pre-K programs. Rather, as with pre-K, the key to success is good design and some minimum level of quality and intensity.  But good design, quality, and time intensity all cost money.  Success does not require the most intense program possible, but it does require some minimum quality level. Early childhood programs do pay off in social benefits, but they also require investments with some significant costs. In the long-run, these costs are paid for with fiscal benefits, but in the short-run there is no free lunch.

Posted in Early childhood program design issues, Early childhood programs | 2 Comments