Does Early Education Reduce the Achievement Gap?

John List and Uri Gneezy have appeared on our blog many times. This guest post is part a series adapted from their new book The Why Axis: Hidden Motives and the Undiscovered Economics of Everyday Life. List also appeared in our recent podcast “How to Raise Money Without Killing a Kitten.”

The past 60 years in the U.S. has seen dramatic policy changes to the public-education system. The ‘50s, ‘60s, and ‘70s saw desegregation and affirmative action, and since the ‘80s there have been efforts to increase school funding, the introduction of voucher systems, and the creation of countless charter schools. In between we’ve seen efforts to reduce class sizes, introduce technology into classrooms, improve teacher credentialing, and a massive attempt to leave No Child Left Behind. 

achievement gap

What do we have to show for all this? That’s hard to say. Even though many programs have a high price tag, they were never implemented with an eye towards assessment. The data we do have shows that not much has changed over the past 30 years. The figure attached shows how the racial achievement gap in test scores has persisted for white and black Americans since the late 1970s. 

If you don’t like the breakdown by race, then consider that the high school dropout rate among high-income families in 1972 was 2% and in 2008 it was still at 2%. For low-income families, though? In 1972 it was 14% and in 2008 it was still at 9%. This sort of trend (or lack thereof) is manifested in dozens of measures of academic achievement, all of which suggest that the past 60 years of educational reform has very little to show for itself.

One of the reasons for our initial interest in educational reform was that we’re really good at testing ideas, but when we first started working with poorer urban schools we found that change was hard to effect. Rewarding parents, teachers, and students for higher grades, better study habits, and improvement on test scores all led to improvements that varied between modest and promising, but none were a silver bullet.

Around the same time a consensus was building in policy making and academic circles that early childhood interventions—like high quality pre-schools—were the most promising prospect for the next round of educational reforms. But testing the impact of an educational reform like that would be unthinkable. It would require millions of dollars, a huge team of teachers and researchers, a school district with an eye towards assessment willing to work with academics, and an army of researchers. 

So that’s exactly what we did.

Just thirty miles south of Chicago we started working with the Chicago Heights school district (which is demographically and economically identical to Chicago) and we found donors willing to invest in our idea in Ken and Anne Griffin. Then we went ahead and built two pre-schools and a parent academy and randomized families into a control group, a pre-school program, or a program that would work with parents to improve their efforts in raising their children.  

How’s it going? Well, the experiment is still in the early stages, but the results so far have been very promising. Students in the two preschools are now doing better than the average child across the nation. These results are astonishing, given the fact that Chicago Heights preschoolers lagged severely behind the average before we started the program.

The children whose parents are in the Parent Academy haven’t seen such stark growth in their test scores, but perhaps most promising is that the growth that did occur persists. Children in the pre-school program seem to have a nasty habit of forgetting what they learn over breaks like summer vacation. 

But this test is still in its early stages. We have an eye towards the long-game here. We’ll be following the families of participants for the rest of their lives, with the goal of informing a whole new generation of educational reform.

If you want to explore our world further, take the Why Axis Challenge: visit, post a photo of your copy of The Why Axis, and be entered to win prizes, including a meeting with Uri, John and Freakonomics author Steven Levitt! 

Leave A Comment

Comments are moderated and generally will be posted if they are on-topic and not abusive.



View All Comments »
  1. Duane Swacker says:

    Hidden due to low comment rating. Click here to see.

    Disliked! Like or Dislike: Thumb up 1 Thumb down 11
    • Taylor S. Marks says:

      I think we all know this system is broken. The question isn’t whether it works or not, but what better system we could use, if one could even exist. The current one offers a means for colleges to accept/reject students and do so in a system that seems, at least for most schools, fair (some of the schools which simply receive too many applications get overwhelmed no matter what. Some are honest about it. Others are dicks – IE, “We’re not wasting paper to tell you this” Stanford.)

      Thumb up 0 Thumb down 0
      • Enter your name... says:

        I don’t even agree that the current system is “broken”. It achieves certain goals quite effectively. Standardized tests are probably the best, most efficient way of figuring out whether someone can do simple math, for example, and they’re not actually that bad these days at handling simpler writing exercises.

        The complaints against testing usually have a lot more to do with wanting non-content skills to be favored, like “creativity” or completely non-academic skills, like good emotional skills.

        Thumb up 2 Thumb down 0
  2. Jason says:

    A well-designed experiment certainly cannot hurt. However, there is already a mountain of evidence in favor of early childhood education and its long term effects. The most famous that I can remember is the NC Abecedarian project ( .

    The problem in education is not the presence or absence of evidence but the lack of political will to follow the evidence. We could have invested a few billion dollars (a minute fraction of our current debt) in universal preschool 20 years ago and been much better off than we are today.

    Thumb up 3 Thumb down 3
    • 164 says:

      I would have to respectfully disagree. If there really was a “mountain of evidence” then it would be a done deal. Statements like yours oversimplify the problem and suggest more government funding as the only solution without looking at the big picture to explore other alternatives. That’s the whole point of looking at the hidden side of everything. I think the point of the blog post is to better understand what works and how long-lasting are the improvements.

      Consider for example how research has shown that Headstart improves outcomes initially but the benefits are lost within the first two years.

      Well-loved. Like or Dislike: Thumb up 10 Thumb down 0
      • Phil Persinger says:


        A “mountain of evidence” exists for climate change and evolution, but that hasn’t resulted in a “done deal” for either when it comes to public policy.

        You’re correct that some research indicates students’ back-sliding subsequent to Headstart. Jason clearly agrees with you about this; his complaint concerns the lack of political will to institute comprehensive day-care and pre-K programs (as described in the Freakonomics post and in his comment) and to support the gains made there in an evidence-based manner all the way through high school graduation. This is no more an over-simplication of the situation than your own comment.

        An enterprise of this scale and persistence (interstate highway system, Apollo program, armed forces) seems achievable only through government funding or subsidy– unless you can describe a purely private-sector alternative.

        Thumb up 3 Thumb down 1
      • Taylor S. Marks says:

        Phil (Whose post lacks a Reply button) –

        I would say interstate highways could definitely be done by states alone. States have built plenty of other highways – what is so special about interstates that they couldn’t have worked it out with their neighboring states on their own?

        Further, what on earth would you want a pre-K program to teach? Elementary school is a freaking joke – it amounts to be little more than all day mandatory babysitting by the government.

        Thumb up 1 Thumb down 0
      • Phil Persinger says:


        Thanks for your reply. I too regret the restricted “reply” options here.

        Only a general convention of all the States– to hash out common routes, construction standards and methods for subsidizing costs in low-income, low-population and/or large-area States– would suffice to put together anything like the highway system we have today. Imagine how that would go.

        Please remember that the legislation was overwhelmingly passed by a closely-divided Congress as a national defense measure (as were many laws in those days)– and, not so incidentally, served the promotion of interstate commerce.

        Elementary schooling should not be a joke– whatever your (or my) personal experience with it might have been. The point of the studies cited in this thread is to determine those measures which will improve the quality of education. And “baby-sitting” by the government is precisely what is needed for working-single-parent and multiple-job-lower-income families who have no reasonable alternatives– irrespective of the educational value this service might provide.

        Maybe some large-scale, long-lasting systems can be put together with purely private planning and funding, but the US and the rest of the industrialized nations proceed by public means for certain enterprises because, for better or worse, private means have proven inadequate to those particular tasks. If one stipulates that the military is one such task, for instance, then it follows that other tasks may at one time or another fall into that same category.

        Thumb up 4 Thumb down 1
      • Jason says:

        (I expect this thread is dead, but I’ll add my follow-up anyway.)

        I do not mean to oversimplify or to suggest more government funding as the only solution. In fact, I would say the problem is far from simple. Despite the complexity, we are presented with decades of research pointing toward early childhood programs. Isn’t part of the reason government programs fail that they often do not follow the evidence?

        In climate change, we have a well defined problem, but the solutions are difficult. (Can humans reverse warming at this point? At what cost?)

        With education, we have a poorly defined problem (we want “better” education for kids), but a lot of evidence-based solutions. My original comment was directed toward the fact that we spend a lot of time testing “solutions” to education, but we rarely act on the results of those experiments, even if they succeed. IMO, that is because they are never “perfect” (since the problem is unclear), so we often move on to a new experiment instead of implementing the results of the prior one.

        Thumb up 0 Thumb down 0
      • Phil Persinger says:


        Thanks for your comment. Let’s keep the thread alive…

        The definition of the goal of any common enterprise is fundamental. Unfortunately, goals in matters of public policy tend to shift over time. For instance (and to vastly over-simplify), the aim of education has drifted from the simple preparation of good citizens to the preparation of good workers. Without a stable goal, we have accumulated a mountain of evidence which may only serendipitously apply to the real education of children– whatever that might be.

        My own point in this regard seems to parallel yours: that evidence from the “hard” sciences currently has just as much influence over public policy as evidence from the “soft” social sciences– which is to say “not much.” I would hazard that public policy in the current climate seems determined far more by cultural bias and the profit motive than by close inspection of the natural world and the human condition.

        Your reiterated point about lack of political will is a true assessment of the contours of this landscape. Facts become noise; debate focuses on shiny objects; attention spans are measured in Angstrom units. This has not always been the situation.

        Thumb up 0 Thumb down 0
  3. Thalia says:

    How is a 30% drop in dropout rates considered not a sufficient result? That seems pretty significant from here.

    Well-loved. Like or Dislike: Thumb up 10 Thumb down 0
  4. steve cebalt says:

    Hi Thalia: I agree — a drop from 14% to 9% in dropouts strikes me as very significant, but how do I know? Likewise the chart above also shows what looks like a very significant narrowing of the gap in scores for blacks vs. whites, contrary to the authors’ statement that not much has changed. I’ve noted similar inconsistencies with these authors in other blog articles. I might be interpreting this wrongly, but the authors fail to explain these basic and important things, so I am left to my own feeble efforts.

    Thumb up 3 Thumb down 0
    • Phil Persinger says:


      Thanks for your Mother’s Day essay, by the way….

      The problem with the narrowing of math scores and drop-out rates is perhaps the 30- and 40-year interval involved in the respective instances: that these gaps, however diminished, remain after all the time and treasure invested. It’s a big public policy question which can easily be expanded from education into areas like income inequality and racial discrimination. Many (like Taylor Marks above) seem to have abandoned support of public education out of frustration– and who can blame them?

      Thumb up 2 Thumb down 0
      • steve cebalt says:

        Hi Phil, and thank you! Good point about the long time interval. I live in Indiana, where vouchers are the current experiment in public/private schooling, using tax dollars to allow people earning $128,000 or less to place their kids in private schools. It’s already shaking up the system in a dramatic way. One problem: No one established goals or measurements. So 30-40 years hence, people will be struggling to say whether it worked — only after several generations of kids have been through the experimental new system.

        Thumb up 1 Thumb down 0
  5. Myles says:

    Great project. One quibble – you write that a drop from 14% dropout rate to 9% is “little to show.” But isn’t that hundreds of thousands, if not millions, of children every year? I’m no math wizard, but that sounds like a third fewer dropouts. Are there that many initiatives that can claim such a success rate?

    Thumb up 2 Thumb down 0
  6. miK says:

    Why aren’t Asians represented in racial data? From my experience, when they exclude Asians when discussing any racial disparity, it’s typically propaganda and not factual.

    High schools can (and did) affect changes even with poor (or non-existent) preschool. Garfield high school was a good example. Read “Standing and Deliverying” by Henry Gradillas on how this was done.

    Thumb up 0 Thumb down 0
  7. Steve says:

    How interesting that cultural attitudes of ethnic groups towards education, rural vs. urban locations, one parent households vs. two parent households with Stay at Home moms, have no representation in the article. You know, just because some things are difficult to quantify, or deemed culturally inappropriate by the liberal tenured elites, does not mean they have no impact. You guys seem to be fishing hard for answers, but maybe you are just asking the wrong questions. Are you really trying to say that the elements in your equation for academic success($/student, head start, etc.) bear more importance to a father in the household, and a mother who ensures her kids go to school and prepare for their future? You are missing the forest for the trees.

    Thumb up 1 Thumb down 0