(Illustration by Ben Wiseman)
Using students’ test scores to assess the efficacy of their teachers—a method known as the “value added” (VA) approach—has become increasingly prevalent in recent years. It has also become increasingly controversial. Critics suggest that it’s neither fair nor accurate to base teacher evaluation on the performance of students who bring widely varying skill levels to a classroom. They also question whether having a high-VA teacher is pertinent to students’ long-term well-being. In a pair of linked papers, three researchers—Raj Chetty, John N. Friedman, and Jonah E. Rockoff—investigate both of those criticisms.
In the first paper, the researchers seek to determine whether the VA method captures the true effect that teachers have on students. To test the possibility that high-VA teachers benefit from having better students in their classroom, the research team developed a way to measure the discrete effects that teacher quality and student quality have on test scores. The researchers found that the VA method does correctly adjust for difference among students.
In the second paper the research team tackles the other issue raised by critics: Does higher teacher quality (as measured by VA) have a positive long-term impact on students’ outcomes in adulthood? Or do high VA scores for teachers simply reflect their ability to help students do well on tests? To investigate that question, the researchers used data from a large school district on students who were in grades 3 through 8 between 1989 and 2009. They then linked those data with data from tax records for the years 1996 to 2011. They were able to align about 90 percent of the school data with corresponding tax data, and that combined data set allowed them to track about one million children from elementary school to early adulthood (age 28).
The researchers used two research methods to estimate the long-term effect of teacher quality. In one method, they looked at cross-sectional comparisons between classrooms. In the other, they focused on the effects of teacher turnover. (What happens, for example, when a high-quality teacher is replaced by a lower-quality teacher?) With both approaches, they found that high-VA teachers had a measurable effect on variables such as lifetime earnings, college attendance, and teenage birth rates. For students in a given classroom, for example, a one-standard-deviation improvement in teacher VA raised the probability of college attendance at age 20 by about 0.85 percentage points.
“We find that good teachers are incredibly valuable,” says John N. Friedman, a professor of economics at Brown University. “Replacing an average teacher with a great teacher is the same thing as depositing hundreds of thousands of dollars into the future bank accounts of the children in a class.”
Andy Baxter, vice president for educator effectiveness at the Southern Regional Education Board, calls the researchers’ two papers on teacher quality “hugely pivotal” and says that he cites them frequently. “They show that what we are doing really does matter in tangible and measurable ways that are long lasting,” he says. “The research base is probably as strong as it has ever been in supporting policies like VA and [in showing] the impact of good teaching on student outcomes as adults.”
Baxter notes that this research appears just as a backlash has taken hold against teacher evaluation approaches such as VA. “The public outcry [against focusing on test scores] is probably at its highest right now,” he says. In response to that point, Friedman cites several alternative applications of the VA method. Educators might use it to measure the quality of teacher training programs, for example. Or they might draw on VA techniques to reward outstanding teaching performance. “Traditional VA has been thought of as ‘We are going to fire the bad teachers,’ and I think there’s a role for that,” Friedman says. “But there are a ton of ways that VA can be used both to improve the quality of teaching and to recognize excellence.”
Raj Chetty, John N. Friedman, and Jonah E. Rockoff , “Measuring the Impacts of Teachers I: Evaluating Bias in Teacher Value-Added Estimates” and “Measuring the Impacts of Teachers II: Teacher Value- Added and Student Outcomes in Adulthood,” American Economic Review, 104, September 2014.
Read more stories by Adrienne Day.