How hard is it to achieve quality on a large-scale in pre-K programs?

Several prominent education and political bloggers have recently argued that implementing high-quality pre-K on a large scale is not a “proven” solution, but rather is hard to do. Kevin Carey of Education Sector argued on December 6 that high-quality pre-k is not a “proven “ solution because the research evidence for “robust long-term effects mostly relies on “small localized pre-K programs”.  Sara Mead of Bellwether Education Partners said that “delivering quality public services for children, of any sort, at scale is HARD, and getting to quality at scale in some of these services – particularly health care and child welfare services – is probably a lot harder than fixing the K-12 system.” Matt Yglesias of the Center for American Progress argued that “proving high-quality preschool on a mass scale isn’t some kind of easy to implement alternative to the tricky task of providing high-quality elementary school on a mass scale.”

As these arguments suggest, part of the concern of these bloggers is the politics of K-12 education reforms.  The concern is that groups opposed to some K-12 education reforms may use pre-K or other services to children as an argument against education reforms. If child outcomes can be dramatically improved via large-scale pre-K or other solutions that are more straightforward than K-12 reforms, perhaps these easier solutions should be preferred.

I certainly agree that the quality of pre-K is important. In chapter 5 of my book, Investing in Kids, I show that variations in pre-K quality make an enormous difference in the local economic development benefits of pre-K programs. For example, feasible variations in pre-K class size can alter economic development benefits by over 30 percent. In later posts in this blog, I will explore some of the effects of various features of pre-K quality on program effectiveness.

However, I think that these bloggers have overstated their case. Although there are some challenges to ensuring pre-K quality, the evidence suggests that it is quite doable for a state or school district to achieve quality in pre-K on a large scale, IF it is willing to spend the needed resources per child and adopt reasonable program design standards. Pre-K is a program that your mythical “average American state” or “average American school district” can implement in a quality way on a large scale.

The evidence on the long-term effectiveness of pre-K goes beyond small “hothouse” programs such as the Perry Preschool program in Ypsilanti, Michigan.  Long-term effectiveness is also shown for the Chicago Child-Parent Center program. This was a large scale program run by Chicago Public Schools.

Furthermore, programs that are similar to CPC have been run in a number of states, and have shown short-run effectiveness. William Gormley’s studies of Oklahoma’s near-universal pre-K program suggest that it significantly improves kindergarten readiness. The National Institute for Early Education Research has done some studies of pre-K programs in a variety of states that also show evidence of short-run effectiveness.  These studies all use what is called “regression discontinuity” analysis to measure program effectiveness. “Regression discontinuity” analysis is considered the next best thing to random assignment experimentation in measuring a program’s effects.

These state programs evaluated by Gormley and NIEER are generally run at a large scale, and are run by “ordinary” school districts or other agencies, not in some extraordinary way monitored by researchers.  Of course, this short-run effectiveness does not prove long-run effectiveness. However, the short-term results are favorable enough and similar enough to the CPC program that it is reasonable that these programs will have long-run effects.

It is not surprising that pre-K helps improve child outcomes in a way that can be straightforwardly implemented. Pre-K, if run at appropriate class sizes, and with appropriately skilled teachers and a good curriculum, adds additional learning time. We know that one reform that works in education is adding more “time on task”. Pre-K is one straightforward way of doing so.

None of this logically implies that K-12 education reforms aren’t also needed. It would be a shame if the need for pre-K was used to argue against K-12 education reforms. Indeed, K-12 education reforms may help the effectiveness of pre-K and other early childhood programs. For example, it has been argued by Ellen Galinsky that part of the success of the Abecedarian program (a full-time child care and pre-K program from birth to age 5) was its location in Chapel Hill, which had a good K-12 system that could help follow through on the boost provided by the Abecedarian program.  In sum, high-quality pre-K and K-12 education reforms can complement each other.

About timbartik

Tim Bartik is a senior economist at the Upjohn Institute for Employment Research, a non-profit and non-partisan research organization in Kalamazoo, Michigan. His research specializes in state and local economic development policies and local labor markets.
This entry was posted in Early childhood program design issues, Early childhood programs and tagged . Bookmark the permalink.

9 Responses to How hard is it to achieve quality on a large-scale in pre-K programs?

  1. None of the studies or programs, have proven that a large scale government intervention on early childhood will lessen achievement gaps. The studies conducted by NIEER and Gormley were funded by PEW, the largest advocates of universal pre-k in the nation. This hardly makes them objective.

    What the studies do prove is that we will see the largest return on investment if government funded preschool is prioritized and given to those children most in need.

    According to NIEER, it would cost 68.6 billion to provide preschool to all 3 & 4 year old children. If we were to target that money to the poor, it would only cost 11.6 billion.

    It would more make sense to use the 50 billion difference to fund K-12, special education or early intervention programs, especially in these dire economic times where all of these programs are being cut.

    Contrary to what the advocates claim, there are plenty of high quality preschool programs available. The massive middle class entitlement program advocates are proposing will only benefit the special interests, lobbyists and education unions.

    The true solution to early childhood is to take our limited funds and put it where the need is. Universal Preschool is not solution to the root of the problem, which is poverty.

  2. timbartik says:


    Thank you for your comment. I note that the subject of my post, and the blogs it comments on, is whether pre-k can be run in a high-quality way on a large-scale. You are commenting on a somewhat different issue, whether pre-k should be universal or targeted on the poor. Even if pre-k is targeted on the poor, to reach all those eligible would require a large-scale program.

    I discuss the issue of universal vs. targeted pre-k at some length in chapter 8 of my book Investing in Kids. I end up preferring universal pre-k to targeted pre-k. However, it is certainly true that there are pros and cons of both approaches, and that we would benefit from more evidence.

    Let me make a few brief response to your comment:
    (1) Gormley and NIEER are well-done studies that use a rigorous methodology. These results should be taken seriously.
    (2) We don’t have strong evidence on the long-term benefits of pre-k for the non-disadvantaged. We do know that many upper class families choose to pay for high quality pre-k, which suggests they perceive that there are benefits. Gormley’s results for Oklahoma also suggest that the kindergarten readiness effects of pre-k do not vary much with whether a child is eligible for free and reduced price lunch.
    (3) Universal pre-k does not necessarily mean that the public sector delivers pre-k. I think the current evidence suggests that a variety of delivery approaches can work. Pre-k can be effectively delivered through the public schools, as is done in Oklahoma. Pre-k can also be effectively delivered through publicly-funded private pre-k programs, as is done in Georgia.
    (4) Universal pre-k does not necessarily mean free pre-k for all. Sliding scale fees can be charged based on income.

    In areas that have many existing high-quality private pre-k programs, it makes sense to build on this infrastructure rather than to scrap it. This can be done by providing public funds to private pre-k programs that meet reasonable measures of quality. In addition, if there is concern about excessive middle class entitlements, sliding scale fees can be used, which will save public funds. Both private pre-k programs and sliding scale fees can be part of a universal pre-k program. However, universality makes it easier to encourage income integration of pre-k programs, which may increase positive peer effects. In addition, as I argue at greater length in my book, universal programs are likely to have more consistent public support for high quality.

  3. Targeted vs. universal is extremely relevant to the topic. How hard is it to achieve quality on large scale in pre-k programs? My answer, not likely. Many believe we haven’t achieved quality for K-12 or other government programs like Head start. Why expect we’ll do any better with pre-k? How will we pay for this program? Who defines what high quality is?

    NIEER and Gormley (and you) are advocating to spend large sums of government money on unproven programs. There isn’t any proof supporting the claim that large scale programs will produce long term benefits. This fact also needs to be taken seriously.

    You agree we don’t have strong evidence on the long-term benefits of pre-k for the non-disadvantaged. You state that many upper class families choose to pay for high quality pre-k, which suggests they perceive that there are benefits. So the government should fund a program that costs billions of dollars because of perceived benefits?

    A couple of notes about Oklahoma and Georgia.

    1. In one of Gormley studies he found children that paid full price for lunch didn’t fare very well in the full day program. Their language skills actually decreased by 24.6%.

    2. How effective is Georgia’s publicly-funded private pre-k programs? After 15 years of UPK, Georgia has failed to lessen the achievement gap and their reading scores are below the national average.

    3. We could go back and forth with varying opinions on what the studies actually say.

    I agree, It would make sense to build on existing infrastructure rather than to scrap it. Unfortunately, that is not the case. The reality is that they ARE scraping existing programs. New York State has been trying to implement a UPK program for over 15 years without much success. Why? Because the state can not afford such a program. Selection is done by random lottery, income can not be a consideration. The reality is the preschool population is being shuffled from the privately paid sector to the publicly funded. High quality programs that have been serving children for many, many years are being forced to close their doors. In reality, they aren’t reaching more children, they are just taking away choice.

    You say universality makes it easier to encourage income integration of pre-k programs, which may increase positive peer effects. According to Gormely’s research, there can also be negative peer effects. This was his explanation for the 24.6% decrease in language skills. The decrease is blamed on the “Peer Effect.” “Students from more disadvantaged backgrounds are more likely to be enrolled in full-day programs than students from more advantaged backgrounds “

    You say, “In addition, as I argue at greater length in my book, universal programs are likely to have more consistent public support for high quality“. How can anyone promise high quality when there is no accountability now for teacher performance? I will repeat the questioned I posed before. Who will define what quality is? Regarding cost, who will set the limits? How much is too much? The private sectors serves its client base, when that client perceives they aren’t getting what they pay for they go some where else and that business goes under. In the public sector they just raise taxes to spend more money. NYS spends the most money in the country per student and we land at number 40 in results, and that comes from Andrew Coumo himself.

    Your argument in support of universal programs is the root of the issue. Public support means more money. Due to well funded campaigns, the advocates have convinced the public UPK is the way to go, when in fact nobody can prove this. This campaign is not about what is best for kids, it is about getting large amounts of public dollars to support an unproven program.

  4. timbartik says:


    You raise many points. I think we are going to have to agree to disagree on what the research evidence is, as I doubt we will resolve all these points.

    As I said before, even if pre-k is targeted on the poor, it would have to be large-scale. So, if you don’t think pre-k works at a large scale, you should be opposed to targeted pre-k as well as universal pre-k.

    But I think the evidence from the Chicago Child-Parent Center program suggests pre-k programs can be run at a large scale and get long-term results. And other large-scale programs get short-term results.

    I’m not sure what you mean by “unproven” programs. Pre-k programs have as much research evidence to support effectiveness as any educational and social program that has ever existed. If we are only going to support programs if every last element of the program has shown effects in large-scale random assignment studies 40 years later, I think that is a recipe for never trying any innovation in public policy. Did Social Security have “proven” long-term results from rigorous experimentation when it was adopted?

    I favor holding pre-k programs accountable for results. I outline ideas on how to do so in chapter 6 of my book, “Investing in Kids”. I may discuss this in a future blog post. The basic idea is to make sure consistent data is collected on student learning gains.

    I would be curious if you would support universal pre-k if it included private providers in a prominent role. Is your objection to universal pre-k as such, or to how universal pre-k is being implemented in New York State?

    I should mention that in addition to writing this book, I have been involved in a local effort in Kalamazoo County to implement universal pre-k, called KCReady4s. (See their website at .) This effort is still in the planning phase. However, the design of this universal pre-k program has had active involvement from many private pre-k and child care providers in Kalamazoo County, and we expect many of the additional slots in this program to be in these private sector providers. Because KCReady4s has actively involved private-sector providers, the program has attracted widespread support from the existing private-sector providers. If universal pre-k followed this type of design, would you have a different attitude towards universal pre-k?

    • The people at Pew, NIEER and other special interests groups invest a lot of money in research and books to support the need for government funded preschool, which directly benefits them. This investment helps to widen their market and build their campaign. Yet, when dialogue about the mixed results of the research begins, I am often asked to agree to disagree. You are not the first to say this to me.

      To answer your question, I don’t want government money to run a program I do not believe in. It makes no sense to spend 60 billion dollars when you can spend 10 and solve the problem. The other 50 billion would be better utilized on programs such as special education, K-12, early literacy or college funding.

      The inequalities and achievement gaps in our public schools are not there because there is a lack of quality preschool programs. 80% of the nations 4 year olds already attend preschool. There have been preschools around for decades doing an excellent job of preparing children. What the preschools haven’t provided, the parents have. As your studies show, sadly, these benefits are not sustained and fade out over time.

      The problems in our public schools could be attributed to the fact that K-12 is overburdened and the socioeconomic conditions of our society. There are too many families living in poverty. Those are the families we should be reaching. Why water down opportunities and provide preschool to the masses, when we can prioritize funding to those in need and really make a difference. So, if we must do “large scale” programs, give them to those most in need. This is where the money should go. It’s that simple.

      So, do I agree to disagree? Yes.

  5. timbartik says:


    I say we will need to “agree to disagree” because such disputes are unlikely to be resolved in the comments spaces of a blog. This is not meant to imply that these issues should not be debated, or that they are not able to be resolved in some other way.

    I simply disagree with you on what the research evidence shows. I don’t think that 80% of low-income and middle-income children are enrolled in “high-quality” preschool. And I think that many of the effects of high-quality preschool on “soft skills” do not fade over time, but rather, tend to grow in importance. Finally, as I pointed out in a previous comment , the type of high-quality universal pre-school that I focus on in my book could be achieved at a national cost of about $14 billion, not $60 billion.

    Thank you again for your comments and interest.

    • According to NIEER, 80% of 4 year old children do go to preschool.

      All of the studies addressing “quality” have been conducted on government funded programs, not on private preschools serving the middle class.

      Again I ask, Who defines quality? I’m guessing your definition of “high-quality” are the 10 benchmarks set out by NIEER. Those too can be debated. Not one of those benchmarks are results oriented. They are geared at creating high paid teacher and administrative positions. Research indicates that teacher credentials aren’t always an indicator of quality. It is the relationship between the teacher and the child that is most relevant.

      Are the advocates of UPK instituting educational programs to prepare kids for kindergarten or are they trying to make the education establishment responsible for the child care of working middle class families? Are they looking to put young children into large institutional settings for 6 -12 hours a day or are they looking for 2-3 hour programs? When they conduct their research, they should compare apples to apples.

      You are correct, this issue is much too complex to go back and forth on a comments blog. There should be much more discussion and debate before our government jumps on this bandwagon. Other alternatives should be discussed and considered.

      Thanks for the dialogue and have a happy and healthy holiday season.


  6. timbartik says:


    In the future, I hope to have more posts discussing issues of “what is quality” in pre-k programs, and how they should be designed, in more detail.

    Thanks again for your comments, and happy holidays to you as well.

    Tim Bartik

  7. Pingback: The Heckman Equation and the relative difficulty of human capital investments | investinginkids

Comments are closed.