By the Numbers: Holding Teachers Accountable

Published April 1, 1999

Educators in Arizona, Colorado, Florida, and Ohio are weighing the introduction of a value-added approach to teacher and school performance evaluation already used in Tennessee to analyze how well schools and individual teachers are doing their jobs.

The value-added approach, developed by statistics professor William L. Sanders at the University of Tennessee at Knoxville, considers the performance of individual students over time and measures their actual academic progress versus their expected progress. Sanders sees as the system’s greatest asset its ability to diagnose what works and what doesn’t.

With growing interest nationwide in teacher and school accountability, Sanders has become “one of the hottest names in education circles these days,” according to Stephen Hegarty of the St. Petersburg Times. That distinction doesn’t come from preaching to the choir. While teachers frequently excuse the poor performance of their students by pointing to social factors beyond their control–transience, poverty, parental involvement, family crises, and so on–Sanders’ research tells teachers they need look only so far as their mirrors to find the biggest factor in student learning.

“We’ve heard all the excuses: ‘When the students came to me, they were scoring low,'” Arizona Department of Education Associate Superintendent Billie Orr explained to Hegarty. “Regardless of where they start, we need to give students a year’s worth of education in a year,” he added, noting that Sanders’ system could measure that.

In a 1997 study, Sanders and graduate student June C. Rivers examined the cumulative effect of teachers on students in grades three to five, using student achievement scores in math from the Tennessee Comprehensive Assessment Program for two large metropolitan school systems in Tennessee. They found that students who have poor teachers for several years are likely to perform at a significantly lower level than comparable students who have had several years of good teachers.

But the lower achievement scores of students who have poor teachers don’t bounce back when these students subsequently are assigned to good teachers. Although an effective teacher can produce excellent academic gains for his or her students, the long-term effect of ineffective teachers earlier in a student’s career is to reduce that student’s achievement scores. (See “Bad Teachers Have Devastating Effect on Student Performance,” School Reform News, April 1997.)

“If an ineffective teacher isn’t dealt with, children can be permanently harmed,” said Sanders, which is why he is interested in using his approach as a diagnostic tool to identify teachers who need to improve. In his research study, teacher effectiveness ratings were based on how well students actually scored with a particular teacher compared to how well they were expected to score from their prior achievement levels.

“Each student serves as his or her own control,” said Sanders, explaining his methodology. “We are bringing an entirely different statistical approach to the educational process.”

For more information …

A 14-page summary of the Sanders/Rivers study, “Cumulative and Residual Effects of Teachers on Future Student Achievement,” published by The University of Tennessee Value-Added Research and Assessment Center, is available through PolicyBot. Point your Web browser to, click on the PolicyBot icon, and search for old document #2180301.