Competition in Education: Does it Make a Difference? An exclusive interview with Dr. Paul E. Peterson
Dr. Paul E. Peterson's recent studies of voucher programs in Milwaukee and Cleveland
have propelled him to the forefront of the educational choice debate.
While initial "official" analyses of the Milwaukee voucher program concluded
that the choice students performed no better than those in the Milwaukee Public Schools, a
re-analysis by Peterson and his colleagues showed substantial gains for the low-income
Harvard University's Program on Education Policy and Governance, which Peterson heads,
was formed two years ago and is jointly sponsored by the Taubman Center on State and Local
Government at the Kennedy School and the Center for American Political Studies in the
Department of Government.
As director of PEPG, Peterson is surrounded by scholars who are examining issues of
educational policy and governance from diverse disciplinary perspectives. He was recently
interviewed by School Reform News Managing Editor George Clowes.
Clowes: What led Harvard University to establish the Program on
Education Policy and Governance?
Peterson: The Kennedy School of Government and the Harvard Faculty of
Arts and Sciences have long been interested in looking at questions of school governance
because schools, like other public institutions, are part of our government. Also,
President Rudenstein had called for greater cooperation across departments and units of
the university, so this initiative was in response to his request and to look at
substantively important areas, specifically at education.
Clowes: How did you become involved?
Peterson: When this initiative came up, I was a logical person because
I have been studying school governance for a very long time. I had a joint appointment in
education and political science at the University of Chicago from 1967 to 1982. I have
become particularly interested in the question of school governance in recent years
because of my concern that central city school systems seem not to be able to turn
We need to look not just at what is happening in the classroom, which is what a lot of
educational research does, but we need to look at what our governing arrangements are to
see if they can be made more effective.
Clowes: Why should we change the way we govern our schools?
Peterson: There are at least three different ways in which people
think about school governance questions. One question is whether or not one needs to
create more centralized direction. Some people think that if we set up national standards
and gave more direction from the Department of Education, we could get a more effective
There's another movement that says we need to have equal funding of education
throughout a state. Then there's a third way, which is where we've spent most of our time,
and that is to create more competition in the system. Can we get more effective governance
by having more competition among school systems and school districts, and what are the
promises of that route?
Clowes: So that's a more indirect approach.
Peterson: Yes. Then there's the further question of whether or not
it's a fair, equitable arrangement to declare 10 percent of our schools--schools in the
private sector--and the families who send their children to these schools, as ineligible
for public assistance for the education of their children. They must pay taxes, and yet
they get none of the benefits of the taxes for their children's education.
Clowes: What were your objectives in the first two years of the
Peterson: The program has taken off in a particular direction, and
that is to encourage the use of randomized experiments to examine the efficacy of
particular governing arrangements. Specifically, whether or not school choice can be shown
to be effective if you subject that governing arrangement to the most rigorous examination
One of the things that I have long thought is necessary in educational research is that
we do more randomized experiments. In medical research, this is a matter of routine. It's
extremely difficult to get a drug approved by the FDA unless a randomized experiment has
demonstrated the efficacy of the innovation. Which is to say, you have to have, by chance,
one person gets the treatment and another person doesn't. Then you look at what happens to
these two people, or groups of people, over a period of time to see whether or not the
treatment is effective.
Clowes: And in sufficient numbers, too?
Peterson: That's right--although surprisingly modest numbers. In
medical research, you can get approval for a drug with a relatively small number of cases
so long as you get statistically significant results. For example, the FDA just removed a
drug from the market because they were coming up with pretty significant side effects as a
result of a randomized experiment.
Randomized experiments are very powerful analytic tools, the most powerful ones
available in any kind of scientific research. And yet, it's very seldom done in education.
A lot of educational innovations take place without ever subjecting that innovation to
rigorous evaluatory research.
We were able to participate in a randomized experiment in New York City that's now
under way, and we'll soon be reporting the baseline information from this experiment. The
applicants of this program came from very low-income families where the students had very
low test scores. And they were not altogether happy with their public schools.
Clowes: These are the students in both groups?
Peterson: That's right, all applicants say this. Those that get the
scholarship get it by means of a lottery. Then we're going to compare those who got the
scholarship with those who didn't, and next year we'll be able to give you the results as
to which of these groups performs better on the same tests.
Clowes: Is one year sufficient time to reach a conclusion about
the effectiveness of the program?
Peterson: No, but it'll give us a preliminary indication. We're going
to follow this for at least three years.
We did do something like this in Milwaukee and we found some results after the first
year, but much larger results after four years. However, the Milwaukee study isn't nearly
as good as the New York study because the randomized experiment wasn't under the control
of an evaluation team.
The lottery in New York was conducted by Mathematica, the evaluation team that I'm
working with. Also, in Milwaukee you didn't have the participation of the religious
schools, so it wasn't as widespread a school choice program as what you have in New York
Clowes: I believe another one is being set up in Washington, D.C.,
to start next fall.
Peterson: Yes, and we will be doing exactly the same kind of a study
there that we're doing in New York City. Several thousand students have already applied.
They will be coming in to be tested and the parents will have their income verified
right after the first of the year. Then the lottery will be held this spring and the
program will begin in the fall.
A lot of the resources of PEPG are being devoted to this because a set of events has
occurred that has allowed us to conduct this really high-quality research on a very
important innovation. We have an opportunity to see whether or not school choice in our
central cities can give children from low-income families opportunities to learn that
they're not receiving within the public schools.
Clowes: Doesn't the work of Caroline Hoxby indicate that
competition already exists between school districts?
Peterson: Yes. Her work is really important because she points out
that it's incorrect to say there is not choice in American education today. There's a lot
of choice--the only thing is that the choice comes through picking a place to live.
The amount of choice varies a lot from one part of the country to another. Some
metropolitan areas, like Boston, have a lot of small school districts, each with its own
school board. Miami is the opposite, with one school district for all of Dade County. She
looked at where do kids learn more, where does education cost more, where do schools
emphasize athletics, and where do they emphasize an academic curriculum.
Hoxby finds that where you have more school districts in a metropolitan area, students
learn somewhat more, costs are significantly less, and sports are not as emphasized, and
the academic curriculum is more heavily emphasized.
Clowes: That's somewhat ironic because that's the way the school
districts were before the consolidations in the 1930s and '40s.
Peterson: Yes. She certainly is making a case against consolidated
school districts, at least in metropolitan areas. Her findings say that maybe we don't
have to completely revamp our educational system, we just need to break down these large
school districts into smaller components.
Her evidence suggests that one possible problem with the central cities is that they
are just one big system, and that you could get a lot more competition if you had truly
independent local school boards controlling a very small number of schools. If they
competed with one another, you'd have a system that was much more responsive to the needs
I think that PEPG is really taking this as its central question: How do we create more
choice? Do we want to design this so that we include private religious institutions or do
we exclude them? There is an argument out there that if people go to a religious school,
they're placed in a community of adults that help sustain one another and provide a social
capital to facilitate education.
Clowes: Like a social support system within the schools?
Peterson: Right. It may be that private schools are better able to
create this sense of community, whether they're Catholic or some other religion. That's
definitely something that we're interested in continuing to explore.
We're also interested in exploring the question of whether or not you would undermine
democratic values if you fragmented American education into institutions, especially those
affiliated with religious groups.
Clowes: If you could set one goal for education policy makers,
what would that be?
Peterson: Well, I don't know if there's one, but I'd say there are
three. One, I think we should find some mechanism to hold schools accountable for what we
expect of them. Second, I think we should work for equity. We should try to make sure that
the resources for education are allocated fairly and in such a way as to facilitate equal
educational opportunity. Third, I think we should explore ways of making the system of
education more competitive, so that it's one where you can get increases in productivity
In many other areas of society, we've had absolutely fabulous increases in
productivity. Why can't we get this in the educational sector?
For me the interesting question is how can we do all three of those things, and achieve
a blend and a balance among them.