Liberté, egalité, randomiser! On Obama’s Pre-K Initiative

Esther DufloThe standout policy announcement in President Obama’s State of the Union address was his commitment to implement universal pre-kindergarten education. This is wonderful, but everything depends on how it is implemented. What I want to see is a further commitment from the President to implement this program through a continuous on-going series of randomized trials.

The program appears to draw from thoughts about early intervention synthesized by James Heckman, for example here:

The accident of birth is a principal source of inequality in America today. American society is dividing into skilled and unskilled, and the roots of this division lie in early childhood experiences. Kids born into disadvantaged environments are at much greater risk of being unskilled, having low lifetime earnings, and facing a range of personal and social troubles, including poor health, teen pregnancy, and crime… public policy focused on early interventions can improve these troubling results. Contrary to the views of genetic determinists, experimental evidence shows that intervening early can produce positive and lasting effects on children in disadvantaged families. This evidence is consistent with a large body of non-experimental evidence showing that the absence of supportive family environments harms childhood and adult outcomes. Early interventions can improve cognitive as well as socio-emotional skills. They promote schooling, reduce crime, foster workforce productivity, and reduce teenage pregnancy.

This is a necessary although (of course) not sufficient answer to problems of inequality in the developed world (see alson Brendan’s thoughts here). I agree with Harry Brighouse:

Heckman is consistently accused by lefties of not understanding that poverty, not parenting, is the fundamental problem… but when I read his essay (and hear him talk etc) everything he says is consistent with the (entirely reasonable) assumption that as things stand, though the fundamental problem may well be poverty,… we need to look for policy levers that would improve the prospects of poor children without addressing their poverty.

My primary disappointment with the plan, however, is that I have not seen a commitment to implement the plan through experiments. I am not saying that the plan is not evidence-based. The White House bases its plan not just on Heckman’s theory but also on evidence from experimental trials, such as the Perry Preschool project and the Abecdarian project. Rather, I am arguing that we should implement the pre-K as a system of overlapping randomized trials.

Here’s why. First, the evidence for the efficacy of existing pre-K interventions is encouraging but far from conclusive. Waiting for conclusive evidence would take a generation; so we need to move ahead now and learn as we go. Second, taking small but successful projects like Perry, Abecedarian, or the Harlem Children’s Zone to scale is a difficult scientific challenge on its own. Finally, even if the existing intervention models worked and could be straightforwardly generalized, there is no reason to think that programs such as Perry, Abecedarian, and HCZ have accomplished everything we could accomplish through early intervention. What we want is a learning system that is continuously experimenting to improve functioning.

What I want President Obama to do is to wherever possible implement the Pre-K initiative following the principles of evidence-based medicine and the closely related principles of the experimental approach to development economics, as expounded by Abhijit  Banerjee and Esther Duflo (the woman in the picture at the head of the post). As Banerjee and Duflo note

First, effective policy-making requires making judgments about the efficacy of individual components of programs, without much guidance from a priori knowledge. Second, however, it is also difficult to learn about these individual components from observational (i.e. non-experimental) data.

In medicine and in policy, almost nothing that works in the lab works in the same way in the real world. The solution is to transform the real world into a laboratory. This is not utopian; it is the everyday practice of the high functioning organizations of our time. Google is built as “an infrastructure of overlapping experiments;” the pre-K initiative should be built the same way.

About Bill Gardner

A health care researcher and a child and quantitative psychologist by training. I am an American living in Canada and am Professor of Paediatrics, Obstetrics & Gynaecology, and Community Health & Epidemiology at Dalhousie University; and Professor of Pediatrics, Psychology, and Psychiatry at The Ohio State University. I also blog at The Incidental Economist ( and you can follow me @Bill_Gardner on Twitter.
This entry was posted in Blog posts. Bookmark the permalink.

One Response to Liberté, egalité, randomiser! On Obama’s Pre-K Initiative

  1. # says:

    Vinyl Strip Door SafetyBecause vinyl strip doors is in afreezer or cooler.
    Those don’t qualify Many windows have U-Factors of. The EZ toggles and anchors work great, and guess what? However, at this time is available to potential crooks.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s