Differences in Performance WITHIN Schools: Why So Much Greater Than in Other Countries?

Cross-posted at Education Week.

Yup.  We know where the differences are in school performance.  They are between the rich schools and the schools serving the poor; between the majority majority schools and the majority minority schools.  They are between the schools that can afford to hire the best teachers and the schools that cannot.  They are between the leafy suburbs and the grim inner cities.  In other words, while we know that there are differences in performance within schools, the big differences in student performance, the ones that really count, are between schools.  That’s why parents are willing to spend a lot more to get their children into schools in the leafy suburbs.  That’s why our accountability systems are focused on giving schools letter grades and singling out the poor-performing schools for special attention.

But then there is the graphic in front of me from the OECD titled “Variation in Science Performance Between and Within Schools.”  It tells a very different story.  Out of 68 countries surveyed, between-school variation accounted for 30 percent of differences in student performance, while within-school variation averaged 69 percent.  Hmm.  Maybe what we know is not true.  Among these countries, the variation in science performance is more than twice as much within schools as it is between schools.

Thumbnail image for Thumbnail image for OECDVariation.png

We rush to look at what the data say about the United States.  It turns out that between-school variation accounts for 20 percent of the variation in student performance in the United States, but within-school variation accounts for about 80 percent.  In fact, out of the 68 countries studied, only eight countries have as much within-school variation as the United States.  Not only is what we know to be true not true; it turns out that it is less true in the United States than in most other countries.

This is excellent news!  Why?  Because it is really hard for school authorities to change the distribution of poor people and minorities among schools, for a whole host of reasons.  It should be a lot easier for school authorities to change what happens within the schools for which they are responsible.

But what, exactly, should they do?  What could explain this kind of variation in outcomes within schools?

May I suggest a couple of possibilities?

First, expectations.  The United States has a long and nearly unique history of assigning students to ability groups from the first grade on.  Nothing, I submit, is more likely to beget big differences in student performance than different expectations for student performance.  See my recent blog on special education for a discussion of this hypothesis.

Second, school organization.  The United States is among the world leaders in the amount of time teachers are expected to be standing in front of a class of students teaching.  There is very little time left for them to do anything else.  A recent large-scale international comparative study by Linda Darling-Hammond that NCEE funded titled Empowered Educators makes it clear that the top performers organize their schools very differently.  In those countries, teachers spend less time teaching and much more time working in teams to systematically improve their lessons and the way they are taught.  They are constantly observing each other teaching, in order to critique each other’s teaching or simply to learn something from their colleagues.  They are meeting with other teachers to discuss the problems of individual students whom they all teach and to work as a group both to understand the problems those students face and to collaborate on the development of a plan to address those problems and then on the implementation of that plan.

Teachers who teach in schools organized this way see themselves as dependent on one another in the same way that members of a law firm are dependent on one another.  They will help one another improve their performance and seek help from others for the same purpose.  Because teachers work very closely together every day, they all know a lot about each other’s performance.  Because they see themselves as dependent for their own success on the knowledge and skills of their colleagues, they will work hard to keep colleagues they respect and just as hard to edge the poor performers out.  Because the career ladders in these countries reward competence at ever-higher levels, teachers have a strong incentive to get better and better at the work.

Contrast what I have just described with the typical American school.  The egg crate school, on the one hand, with each teacher in her own classroom and the door shut has neither time nor incentive to collaborate with colleagues.  She is subject to an accountability system that assumes that the achievement of each teacher’s students is uniquely attributable to that teacher and only that teacher for the subject assessed.  On the other hand, in top performing education systems, we have the school organized as a collegium, with teachers spending most of their time not facing students but working in teams to steadily improve every aspect of the school’s functioning.  These are different worlds!

It is surely true that there are schools organized as collegiums in the United States and schools in which teachers don’t collaborate much in top performing countries.  But the exceptions prove the rule.  It is a striking contrast.  More accurately, it is a devastating contrast.  Is it any wonder that we see far more within-school variation in student performance in the United States than in the world’s top-performing education systems?  Consider the difference in incentives and supports in these two forms of school organization.

In the conventional American school, it is up to the school principal alone to get rid of poor performers and up to the union to protect them.  Other teachers have no reason to get involved.  New teachers, fresh out of teachers college, get very little support from the veteran teachers and often leave the occupation early feeling they have failed at teaching.  Strong teachers survive and do a great job for their students, but others carry on for years, meeting their obligations but doing little more, ultimately just serving time until their pension maxes out.  Each teacher goes his or her own way, like atoms in a lonely universe.

But in a school organized as I have described, new teachers get very strong support and teachers who still struggle get a lot of help every day and every week from highly skilled colleagues from whom they stand to learn a lot.  Good teachers get better.  Better teachers become great.  The weak who don’t get better are the ones who would become time-servers in a typical American school, but, in the top-performing systems, are counseled out by their colleagues.

Perhaps you can see now why I believe that the big difference between the U.S. and the top performers with respect to within-school differences in student performance is largely explained by differences in expectations for students and in the way schools are organized.

No, I don’t think that we can close the whole gap in student performance between the rich suburbs and poverty-stricken inner city and poor rural schools by having the latter adopt the form of school organization I have described.  But I am convinced that we could substantially reduce the gap that way.