John Friedman on the Economics of Kindergarten and Public Policy

Interviewed by Doug Gavel on August 3, 2010

A child’s success in her first year in the classroom can portend her chances of success as an adult. That is one key finding in a new research study co-authored by Harvard Kennedy School assistant professor John Friedman. Along with several colleagues, he examined the records of 12,000 school children who were subjects of an education experiment in Tennessee in the 1980’s, in order to determine how one’s kindergarten education affects her prospects later in life. Friedman holds a Ph.D. in Economics, an A.M. in Statistics, and a B.A. in Economics, all from Harvard University, and his academic research focuses on the economics and political economy of public policy.

Q: Please explain the parameters of the original Tennessee experiment and the data you drew upon for your research.

Friedman:
In 1985 the government of Tennessee started a program in which a group of students entering kindergarten were randomly placed, some to large classrooms and some to small classrooms. Teachers were also randomly assigned to different classrooms, and then as a byproduct of this, the peers that a student had in each classroom were randomly assigned.

The experiment continued for four years through the third grade. For the past 20 years, this random assignment of kids to classrooms has been a vital tool in helping educational researchers understand the impact of teachers, of peers, and of class size on educational outcomes.

This data set followed the students after the program ended in third grade, through eighth grade, and test scores were collected each year. So prior to our research, we could examine the impact of smaller kindergarten classrooms, or that of having a more experienced kindergarten teacher, or that of having a smarter set of peers in kindergarten on a student’s test scores through the eighth grade. But the data stopped at that point, so the question remained – did the impact on test scores through the eighth grade accurately represent the longer term impact of these early education interventions?

Our study sought to address the outcomes that we really care about – whether the child went to college, the quality of that college, whether she is in a high-paying job, whether she owns a home, and whether she is saving for retirement.

Q: Please discuss the most compelling findings from your analysis.

Friedman:
The most significant results from our study are twofold.

First of all, we were able to show that students who were in a smaller kindergarten class were about two percentage points more likely to attend college than those who had been in larger classes; they earned a little bit more salary; they were more likely to be married and more likely to own a home. Also, students with more experienced teachers earned $900 dollars more per year at age 27 than their peers who had less experienced kindergarten teachers. And that gap seems to grow over time, so that if we followed the students for a few more years, that gap might be even larger.

The second major finding of our study responded to the question: Are test scores a good measure of the longer-term impacts of a better teacher or a smaller classroom? What we found was very interesting. We found that kindergarten test scores are a very accurate measure of the impact of a kindergarten class on later life outcomes. For every percentile higher that the students scored in kindergarten, for instance, they earned about $80 more per year later in life. Later test scores were not as accurate a measure. For instance, the impact of small classes is to increase kindergarten test scores, but it has almost no impact on eighth grade test scores. So there’s a conflict there – is the immediate impact upon kindergarten test scores or the long-term impact on eighth grade test scores more relevant for long term outcomes? And it looks like it was really the kindergarten score that was the best measure. The fact that effects faded over time in regards to test scores was not a good measure of the long term impact on outcomes.

Q: Did the findings surprise you?

Friedman:
People tend to have different perspectives about the issues that we studied. Some people, seeing the impact of test scores fading over time, thought that what happened in kindergarten just could not have that much of an impact on one’s life 20 or 30 years later. But other people, especially those who believe that early childhood interventions are really important, felt that kindergarten might really have a tremendous impact later on. So I think that the community was pretty much split on whether our results were consistent with what they had believed or differed from what they believed.

One thing that we can be sure of, however, is that we are not identifying all of the factors that influence a person’s development. Kindergarten is very important, but prior research suggests that the impact of one’s parents is at least three times more important than one’s kindergarten class. Your peers are crucially important, also, not only in class but outside of class as well. So your kindergarten class is tremendously important; the stakes are very high in kindergarten. But just because you don’t have the best kindergarten class doesn’t mean that you are doomed for life; there are other important influences as well.

Q: What are the implications of this research on early childhood education?

Friedman:
I would identify two significant implications. First of all, I think this research validates a tremendous amount of literature on early childhood interventions, because that literature is only able to look at test scores. These findings show that those test scores are indeed a fairly reliable predictor for what goes on in the longer term.

The second major impact of our study, I think, is to highlight how important kindergarten is. The stakes are tremendously high. And this research highlights how critically important it is for educators to support kindergarten and to spend its resources correctly. We have calculated that the difference between an above average teacher and a below average teacher on lifetime earnings of students is about $320,000 for an average size kindergarten class. That’s based on a calculation in which lifetime earnings are $16,000 more per students, and there are 20 students in a class.

I’m not an expert in teacher recruitment, and I don’t have a specific policy recommendation to raise the quality of teachers, but I find it disturbing when states are reducing funding for these types of early education programs, such as Head Start or direct funding to schools. I think the results from our research suggest that that’s not the best way to look after the long term well being of our children.

Q: How do you see these findings as influencing policy makers?

Friedman:
These findings highlight how important kindergarten is for determining the long term success of children. And one of the important sub-findings of our research is that these interventions seem to raise the long term earnings and the probability of attending college just as much for richer students as for poorer students. So this suggests that one way to narrow economic inequality is to focus resources on those students who have fewer opportunities or fewer resources available to them when they are young.

Unfortunately much of the funding for schools in this country comes from local taxes, which means that students who grow up in poor neighborhoods often attend very poorly- funded schools. That might serve to perpetuate the patterns of inequality if kindergarten funding and the quality of the kindergarten classroom are important over the long run. So I hope that in the broader picture our research might help policymakers understand the importance of redirecting some of the early education funding towards those lower income neighborhoods, which, as a result, would serve to equalize the early education opportunities that different students are getting.

Q: What future research projects are you planning to follow up on this study?

Friedman:
I think that the current study is very good at picking out a few specific aspects of the kindergarten classroom that seem to matter and highlighting the overall importance of kindergarten, but we do not make many specific recommendations about how to make a kindergarten class better. We know that small classes are good, and we know that teacher experience seems to matter, but beyond that we are limited in the lessons that we can draw.

So I think the research project going forward will continue to examine what makes a kindergarten class excellent. This research really gives us a lot more power to see, for instance, whether and how peers matter. We have very little to say about the quality of one’s peers, but educational research has time and again found that one’s peers are extremely important. Similarly, it would be nice to know something more about what beyond teacher experience makes a difference – maybe if the teachers have children themselves, or if they’re younger or older, or if they have more experience in that particular age range rather than in schools in general. So I think that is the challenge moving forward – to make specific policy recommendations beyond the more broad brush of the current findings.


John F. Kennedy School of Government 79 John F. Kennedy Street
Cambridge, MA 02138
617-495-1100 Get Directions Visit Contact Page