Sunday, December 5, 2010

Are teachers not part of the learning equation?

Previously published in the Terre Haute Tribune Star, 12/5/10

TERRE HAUTE — “A teacher’s influence on student achievement scores is 20 times greater than any other variable — including class size and student poverty.”

This is a quote taken from Gov. Mitch Daniels’ “Legislative Priorities” ( It didn’t read to me as something he or even one of his staffers wrote. So, I donned my sleuthing persona, and quickly found the quote repeated in many sources, the Houston Chronicle, some foundations, eventually though I traced the quote to Kati Haycock, the director of the Education Trust, an education policy advocacy group.

The Education Trust is a fairly mainstream group. They are funded by such groups as the Bill and Melinda Gates Foundation, the Pew Charitable Trust, and others. I suspect Kati Haycock would smile at some of Indiana’s proposed education reforms and frown at others.

Haycock is not a statistician. Among other things, she edits an in-house journal. She, I think, is making a rhetorical claim, based upon the research of William L. Sanders, who is a skilled statistician/econometrician from Tennessee, who created the Tennessee Valued Added Assessment System (TVAAS). This is an assessment and accountability system ushered into Tennessee under Gov. Lamar Alexander in the mid-1980s. In fact, the model underlying the assessment system is known as the “Sanders Model.”

Many of the claims being made by Gov. Daniels and Education Superintendent Bennett stem, I think, from Sanders’ claims. And he makes some pretty remarkable, albeit controversial, claims. Among them is that class size, heterogeneity of the classroom, social class and other commonly found influences in other research are not important. It is as though once students enter the classroom only the influence of the teacher matters. What I find interesting is that there is a lot of sophisticated research using data from Texas, Florida, North Carolina, Alabama and Kentucky.

Tennessee may have the best data, but I can only find research using TVAAS data by Sanders. Essentially, the creator of the system is the only one who has evaluated the system. (Perhaps the databases I use, which located Sanders’ articles, just don’t find articles published by others using TVAAS data.) Maybe Tennessee won’t grant access to any other researchers? Sanders’ claims cannot be tested by other scholars if the data is not made available.

The value-added model of assessment makes sense for students. To use the same model to assess teachers raises some questions. Any occupation can be divided into quintiles based on productivity, job satisfaction or height. No matter how well the bottom quintile (20 percent) does, they are still low-performing relative to the top. And if you eliminate the bottom quintile and redistribute, there will still be a lowest quintile.

What we need to do, if we are going to adopt value-added assessment to evaluate teachers, is to establish a minimum acceptable absolute performance, not a relative one.

There are other troubling findings in this literature. There is no observable teacher characteristic that consistently correlates with student outcomes. It doesn’t matter the quality of the school a teacher attended, it doesn’t matter how well they scored on any teacher licensure exam, their own SATs, how experienced they are, whether they have a master’s degree, nothing. The idea that good teachers are born, not made, is the inevitable conclusion.

They burn out quickly, however. Research consistently shows that except for the first 10 to 15 years, a teacher’s experience is generally negatively related to student achievement, especially in math.

Superintendent Bennett uses sports metaphors all the time. The research on teacher experience and student achievement suggests this one: Teachers are like NFL quarterbacks. They improve through the first few years, hit a plateau, and then decline. Time to draft a new one.

All of Tennessee’s students take the ACT. In 2010, not one of the areas ACT covers shows Tennessee’s students exceeding national averages in readiness for college. Fewer students in Indiana take the ACT, but Indiana’s students exceed the national averages in readiness for college in each category.

There is not a lot of research on home-schooled students. What research I found shows that home-schooled students do better on standardized achievement tests than formally schooled students. Who home schools? Married couple families with one or two children, highly educated, single-earner families in the mid-$70,000s, and either Mom or Dad stay home to exclusively teach their kids. Interestingly, those parents who happen to also be certified teachers don’t do as well as the students of non-certified teachers. So maybe teachers are not the most important factor; perhaps it is parents’ emphasis on education, small class size, and highly tailored curricula.

Note: Steiger is married to a public school teacher.

No comments:

Blog Directory - Blogged The Steiger Counter at Blogged