STOP THE PRESSES x 2! (From tomorrow's NYT) This is REALLY important research that the Gates Foundation is funding – kudos! (And kudos to Randi and the NEA for allowing the experiment, despite reservations.)

In most American schools, teachers are evaluated by principals or other administrators who drop in for occasional classroom visits and fill out forms to rate their performance.

The result? More than 9 out of 10 teachers get top marks, according to a prominent study last year by the New Teacher Project, a nonprofit group focusing on improving teacher quality.

Now Bill Gates, who in recent years has turned his attention and considerable fortune to improving American education, is investing $335 million through his foundation to overhaul the personnel departments of several big school systems. A big chunk of that money is financing research by dozens of social scientists and thousands of teachers to develop a better system for evaluating classroom instruction.

The effort will have enormous consequences for the movement to hold schools and educators more accountable for student achievement.

Twenty states are overhauling their teacher-evaluation systems, partly to fulfill plans set in motion by a $4 billion federal grant competition, and they are eagerly awaiting the research results.

For teachers, the findings could mean more scrutiny. But they may also provide more specific guidance about what is expected of the teachers in the classroom if new experiments with other measures are adopted — including tests that gauge teachers' mastery of their subjects, surveys that ask students about the learning environments in their classes and digital videos of teachers' lessons, scored by experts.

"It's huge," said Deborah Loewenberg Ball, dean of the University of Michigan School of Education. "They're trying to do something nobody's done before, and do it very quickly."

…The meticulous scoring of videotaped lessons for this project is unfolding on a scale never undertaken in educational research, said Catherine A. McClellan, a director for the Educational Testing Service who is overseeing the process.

By next June, researchers will have about 24,000 videotaped lessons. Because some must be scored using more than one protocol, the research will eventually involve reviewing some 64,000 hours of classroom video. Early next year, Dr. McClellan expects to recruit hundreds of educators and train them to score lessons.

The goal is to help researchers look for possible correlations between certain teaching practices and high student achievement, measured by value-added scores. Thomas J. Kane, a Harvard economist who is leading the research, is scheduled to announce some preliminary results in Washington next Friday. More definitive conclusions are expected in about a year.

The effort has also become a large-scale field trial of using classroom video, to help teachers improve and to evaluate them remotely.

…In addition to the cost — which many struggling districts may consider too high — another barrier could be teacher opposition. The Memphis teachers union, an affiliate of the National Education Association, has partnered with the foundation for the project. But Keith Harris, its president, said the use of videotaped observations in evaluations raised troubling questions.

"Whose eyes would see these videos?" Mr. Harris asked. "Who would own them? This seems like an 'I gotcha' kind of thing. We think these observations deserve a human being."

Randi Weingarten, president of the American Federation of Teachers, which has several affiliates participating in the research, also expressed reservations. "Videotaped observations have their role but shouldn't be used to substitute for in-person observations to evaluate teachers," Ms. Weingarten said. "It would be hard to justify ratings by outsiders watching videotapes at a remote location who never visited the classroom and couldn't see for themselves a teacher's interaction and relationship with students."

-------------------