Emphasize student learning and growth, yes, but don't judge teachers solely on standardized test scores. Make classroom observations count, and train evaluators and teachers so that both have a common understanding of the specifics the observer is looking for. Use varied measures of student learning, including, perhaps, student portfolios, internal assessments and checks, and teachers' goals for their students. Even – crazy as it might sound – ask students about their teachers.
What's important "is ensuring that each of those measures contributes something fairly unique and different to the picture," says Vicki Phillips, director of the Education, College Ready United States Program for the Bill & Melinda Gates Foundation. The foundation has invested $290 million in teacher evaluation programs in three urban districts, including Hillsborough County, and five Los Angeles charter-school networks.
The long-term, ongoing Gates foundation Measures of Effective Teaching (MET) study has in recent years delivered some of the most thorough research on teacher evaluations. It found that the most important elements are detailed observations tied to a specific rubric; student achievement growth, measured, in most cases, by value-added test scores that try to isolate the teacher's contribution to students' scores; and student surveys rating teachers on such factors as how well they support, or challenge, or provide feedback.
"When you add them together, the corroboration becomes even stronger … and you get a much more reliable picture," says Ms. Phillips.
Policy reform brings psychological shift
After a peer evaluator watched Newman teach three-digit addition and subtraction problems to her second-graders last year, the evaluator noted that Newman could do a better job asking her students to evaluate their own work, and suggested she do more to encourage them to reflect on and analyze their learning.