Most teachers have long known that they affect the life chances of children. But it took the work of economists to convince the world of public policy to take seriously what is now known as "teacher effectiveness."

Now one of those very same economists has turned to another subject that, to most teachers and principals, is similarly self-evident: Principals, like teachers, affect the life chances of children, too.

Last week, Stanford's Eric Hanushek -- who conducted many of the early economic analyses on teacher impact -- presented a new research paper at a conference in Washington, D.C., hosted by the National Center for Analysis of Longitudinal Data in Educational Research. The findings show, in his words, that "principals matter."

How much they matter to kids still needs further study and depends on how you run the analysis. Do you control for the kind of schools principals lead? Do you control for how long they've been working in the school? On and on the methodological questions run, but no matter how Hanushek examined the numbers, he found that principals demonstrably affected student achievement.

In some ways, Hanushek is not saying anything new. Others, including Ken Leithwood at the University of Toronto, have shown this in large-scale studies, too. But education professors are easily and often ignored by policy makers. Hanushek is harder to disregard because his body of research is so intertwined with current policy discussions, particularly on teacher impact.

So if principals matter, it would be worthwhile to understand what highly expert and effective principals do that matters so much. Are they good at hiring and keeping strong teachers? Do they structure the work in such a way that ordinary teachers can improve their practice and be successful? Do they establish a climate and culture that encourages teachers to try new things, but ensure that those practices that aren't successful in improving student achievement are not continued? Do they set up systems that allow teachers to focus on the work of instruction instead of having to invent new solutions to every single problem that crops up on campus -- like what to do when the school ran out of paper halfway through the semester?

More broadly, do they establish the expectation that all children will be successful, and then engage all the adults in a school to be part of solving the problems that could thwart such an expectation from being realized?

When my colleague Christina Theokas and I studied the practices and beliefs of 33 high-performing principals who work in schools with substantial populations of low-income students and students of color, we found all of the above. The patterns were remarkably consistent even though each of the school leaders was immersed in idiosyncratic school situations: some rural, some suburban, some urban; some elementary, some secondary; some embedded in helpful districts, some considered pariahs within their districts. In general, these schools are not expected to perform well on standard achievement measures because of the demographics of their students. And yet, they all perform about as well -- or better -- than do white, middle-class schools. These are schools worth paying attention to, and can serve as an example for anyone interested in helping all students learn.

One of the comments during Hanushek's session was from an economist who said that, generally speaking, his field always assumes that managers are interchangeable. Maybe it is time to see that those who manage -- that is, lead -- schools are far from interchangeable. In fact, they are pivotal to our schools functioning as networks of opportunity for all children, and it is worth spending some time figuring out what they do that others can emulate.