The latest results of the massive Measures of Effective Teaching Project may give pause to districts working to develop teacher-effectiveness evaluations.
The Bill & Melinda Gates Foundation's MET project, one of the largest instructional-observation studies in the country, has found that teacher-effectiveness assessments similar to those used in some district value-added systems aren't good at showing which differences are important between the most- and least-effective educators, and often totally misunderstand the "messy middle" that most teachers occupy.
"The beauty of multiple measures isn't that there are more of
them—more can be more confusing. These need to be aligned to the
outcomes we care about," said Steve Cantrell, who oversees the MET
project for the Seattle-based Gates. (The foundation also helps support
coverage of K-12 business and innovation in Education Week.)
Mr. Cantrell and other MET researchers dug into the latest wave of findings from the study of more than 3,000 classes on April 14 for a standing-room-only ballroom crowd at the American Educational Research Association's annual conference here.
Existing teacher-evaluation systems often use too few classroom observations, indicators that are not effective at gauging student achievement, and evaluation methods that lump teachers into too-simplistic categories, the researchers found.
"The middle is a lot messier than a lot of state policies would lead us to believe," Mr. Cantrell said. "Based on the practice data, if I look at the quartiles, all that separates the 25th and 75th on a class [observation] instrument is .68—less than 10 percent of the scale distribution. In a lot of systems, the 75th percentile teacher is considered a leader and the 25th percentile considered a laggard."
Moreover, Mr. Cantrell said, firing the lowest-performing quarter of teachers wouldn't improve teachers' instruction generally. Researchers found differences in classroom practices between effective and ineffective teachers mostly in classroom management and behavior. Based on observing more than 24,000 lessons, the project found that "classroom practice could be described as orderly but unambitious," Mr. Cantrell said.
Is it possible to test whether a teacher will use the best instructional practices, as opposed to only average ones?
Drew Gitomer, the education chair at the Rutgers Graduate School of Education, in New Brunswick, N.J., thinks so. At the conference, he described his work with MET to develop better teacher content assessments, which gauge "whether teachers understand how to phrase questions and examples that will address potential student misconceptions about the content."
For example, one question asks teachers to determine which of several examples of simple exponents would be the least useful in a math class; the question looks for the teacher to understand that 3 cubed would be a better example than 2 squared, because the latter would give students the same answer, 4, if they simply multiply the number by its exponent. Three cubed, if solved incorrectly, would reveal a student's misunderstanding of the procedure.
Mr. Gitomer conducted 90-minute interviews with 60 teachers who scored in the top and bottom 25 percent of teachers on that test. He found that the lowest-performing teachers often had weak reasoning for instructional decisions; they lost track of the larger purpose behind a lesson and often cited personal preference to justify an approach.
Strong teachers, by contrast, used questions to look at larger classes of problems, and could describe how their approach supported their students' learning.
Student observations may be the real key to identifying what works in teaching, said another presenter, Ronald F. Ferguson, a senior lecturer in education and public policy at Harvard University. He analyzed surveys from 2,985 MET classes with at least five responding students each, and compared students' achievement with their observations of their teachers' use of "seven C's" of teaching practice. They focus on whether a teacher:
• Cares about students;
• Captivates them by showing learning is relevant;
• Confers with students to show their ideas are welcomed and respected;
• Clarifies lessons so knowledge seems feasible;
• Consolidates knowledge so lessons are connected and integrated;
• Controls behavior so students stay on task; and
• Challenges students to achieve.
Educators often think caring is the most critical practice for student achievement, but Mr. Ferguson said he found other practices more correlated with high achievement.
Teachers who were in the top quartile in the number of students reporting that they practiced the seven C's, Mr. Ferguson said, had students who achieved a full semester's worth of learning more than students of teachers in the bottom quartile of the seven C's