No magic bullet for evaluating teachers

Brandon Curl


Last week at a town hall at Bell Multicultural High School in Washington, D.C., President Barack Obama decried the use of standardized testing. “Too often what we have been doing is using these tests to punish students or to, in some cases, punish schools,” Obama said. He went on to comment on how standardized testing forces teachers into “teaching to the test.”

What’s confusing about these remarks is not that they are untrue but that they are at odds with his administration’s own policies, which reward federal education funding to states that institute reforms tying teacher evaluations to standardized test scores.

For an example of these efforts, look no further than Thursday’s announcement that the Texas Education Agency (TEA) has contracted with UT’s Lyndon B. Johnson School of Public Affairs to develop a new metric to measure teacher effectiveness. That venture, which falls under the direction of the school’s Project on Educator Effectiveness and Quality (PEEQ), will include analysis of “student performance on state standardized exams.”

Why is Texas doing this? According to the University’s press release, “federal grant requirements for measuring educator effectiveness” necessitate the creation of a student achievement standard, which, in Texas’ case will be the PEEQ metric. In other words: money.

In a year in which Texas is facing a much-publicized multibillion dollar budget shortfall, of which education takes a great share, any available federal funding is vital. Unfortunately, depending whether Obama is right (or wrong depending on your interpretation of his views) about standardized testing, this support may come at the expense of public education in Texas.

The real problem with standardized testing is not its existence but in the insistence on the primacy of its usage in evaluating teacher effectiveness. That point is best exemplified in the conversation surrounding the newest method of evaluating teacher effectiveness, value-added modeling.

Value-added modeling began as a way to improve the way we look at standardized test scores. Traditionally, schools have rated the effectiveness of a teacher by comparing the performance of his or her students to some sort of national average. But with value-added measures, students are compared to themselves. Specifically, statisticians use a student’s past test scores to predict future test scores.
Then the student’s actual test score is compared to that prediction in order to calculate the impact or “value” the teacher has added.

It seems like a good idea, and value-added measures have been championed by the Measures of Effective Teaching (MET) project, a multi-year study funded by the Bill & Melinda Gates Foundation. But the conclusions drawn from the preliminary report released this past December have been widely criticized, most notably by respected Berkeley economist Jesse Rothstein. Among the criticisms to value-added measures are the ideas that the same score gains might not be equivalent for high and low-achieving students and the effects of principal quality are not considered.

Ultimately, standardized test scores should be used as a small piece in a larger, more comprehensive method of evaluating teacher effectiveness that includes performance-based assessments, classroom observations, student surveys and teacher-reflection among other variables. Unfortunately, it is unlikely that we will be able to effectively distill this down to one number as PEEQ has been tasked to do by the TEA.

As much as we may hope for some magic bullet that will efficiently rank our nation’s teachers from least to most effective, we will have to go about it the hard way, and that means qualitatively evaluating teacher effectiveness from every angle.