Expert urges caution when using standardized tests to evaluate teachers

Wesleyan professor says there's no research to prove it works

Testifying before the Education Committee on February 21, Daniel Long, a professor at Wesleyan University, said the state should exercise caution in moving ahead with a teacher evaluation system that uses student test scores. "There have been almost no studies of the long-term effects of actual teacher evaluation programs that use students' tests," he said.

"Claims about the benefits of using student tests to evaluate teacher programs are based on hypothetical statistical calculations that fail to consider the negative effects of teaching to the test, the narrowing of the curriculum, the decline in teacher cooperation, and the decline in teacher morale," Long explained.

If the state plans to use student standardized test scores to evaluate teachers, Long recommends that, at a minimum, "The state needs to run pilot programs for several years in rich and poor towns in Connecticut to evaluate a new teacher evaluation program."

Asked what does lead to a high-quality teacher evaluation system, Long responded that serious training for evaluators is key.

"If we look back at Connecticut's history," he said, "if we just look at what happened in the 90's in education reform, there seemed to be teacher evaluation models that caused a notable increase in Connecticut's performance."

In 1998, Connecticut was the top-scoring state on the fourth grade NAEP Reading Assessment and the state which demonstrated the greatest amount of growth from 1992 to 1998. Connecticut was also the top-performing state in the nation in writing, and was one of only two states to receive three "gold stars" from the National Goals Panel in 1998 for achievement in math and science.

Long said that "An evaluation of Connecticut's achievement done by the National Goals Panel in 1999 exploring high-improving reading achievement in Connecticut found that the nature of high levels of professional development, raising salaries, had a dramatic effect- increasing overall reading scores, and took Connecticut from being a high scorer to pretty much the top scorer in reading achievement on the National Assessment of Education Progress."

However, even the best teacher evaluation system isn't going to be a cure-all for the education issues facing the state, Long continued, "I think the real problem is the wealth gap that exists."

Long said that the wealth gap and the inequalities that go along with it need to be addressed by comprehensive wraparound services from a very early age—before students ever enter school—when the achievement gap begins. "Those are the things that I think make the biggest difference," he said.

"There is a lot of research on the effects of peers on student achievement," Long added. Research shows that in schools that integrate students from a range of income backgrounds, higher socioeconomic status students help raise the achievement of all students.

"Also, there's a feedback effect," Long said. "In the neighborhood that I live in, a working-class neighborhood in Middletown, it's because of the activism of a number of highly educated parents that happen to live in this working class neighborhood, that they're able to dramatically improve the schools."