Tuesday, October 14, 2014

Writer discusses problems with using students' test scores to grade teachers

Districts evaluating teachers based on students' test scores is becoming a more and more common practice. In Houston, the teachers' union even sued their school district as a result. Those who condone the practice often cite research that highlights a connection among teachers, test scores and students' success in adulthood. However, Jesse Rothstein, an economist at the University of California, argues that this might not be a consistent measure of teacher effectiveness because teachers who are viewed as exceptionally skilled may just be "assigned good students or seek them out rather than making a difference in the test scores themselves," Max Ehrenfreund writes for The Washington Post.

Economist Raj Chetty of Harvard University and two colleagues conducted the study in question. They found that students who performed well on tests earned more money later in life. President Obama even talked about the study in 2012 when he said, "We know a good teacher can increase the lifetime income of a classroom by over $250,000."

This information becomes less meaningful if certain teachers are simply assigned or seek out better students. For example, what if one teacher is adept at working with students who don't speak English, while a second teacher has poor classroom management skills. The second teacher may be assigned well-behaved, dedicated students. The second teacher's students will likely perform better on tests—and earn more money in the course of their lives, like the study predicts—but this isn't because the second teacher is better.

While Harvard's John Friedman did find that "when a teacher whose students did well on tests moved to a different school . . . the average score across that teacher's grade at the new school did improve, indicating that it wasn't just a matter of schools assigning the best students to particular teachers, Rothstein wasn't convinced," Ehrenfreund writes. Rothstein observed a similar correlation in test scores and incomes in North Carolina but noted that when a teacher whose students get good test scores moves to another school where scores were already going up, it's hard to determine how much of the improvement can be attributed to the new teacher.

Chetty agrees that "whether value-added scores are the best way to assess teachers is still an open question," Ehrenfreund writes. Also, using students' test scores as the basis for awarding teachers better salaries or tenure might have negative effects if teachers don't seek to cultivate qualities other than academics in their students. "Once you start using value-added measures in practice, their signal quality might get eroded," he said. "People might start teaching to the test." He doesn't think teachers will make that mistake but doesn't have data to support it yet.

Rothstein said he isn't saying the models can never be useful. "I'm arguing that we need to go in with our eyes open." (Read more)

No comments: