Across the country, policies to link teachers to student performance are on a fast track. Spurred by the competition for federal Race to the Top and Teacher Incentive Grant dollars, several states changed laws and regulations that govern the use of teacher-student data in order to put themselves in a favorable position to win the grants. At the same time, numerous urban districts are at work with teachers to implement pay for performance plans, often with the aid of sizable private foundation support.
The focus on teachers is well-placed. Of all the factors that influence student learning, research shows that teachers are the most significant. Yet, with some notable exceptions, teachers havent exactly embraced the idea of being judged by the performance of their students. Many arent convinced that the data is valid, and the effort remains controversial.
Two recent reports shed valuable light on the teacher-student link and show how complex the task of teacher evaluation is. The sources themselves are interesting — One report was published by the Gates Foundation and the other by Center for American Progress. Both organizations are high-profile advocates for using value-added measures of student growth to identify and reward effective teachers as the key strategy in school improvement plans. Both reports see a lot of value in value-added, especially when used for school and staff improvement. But while neither report flat out says don’t do it they also advise care when using student performance for high stakes decisions about teachers.
The Gates study reports the preliminary findings from their multi-year “Measuring Effective Teaching” project. The executive summary is here. Among the more interesting findings: classroom observations and student perceptions have strong correlations to teacher value-added data. If this relationship holds, this could go a long way toward building up teachers’ trust in the data as districts move forward with teacher-student links.
The Center American Progress report is by Dan Goldhaber a widely respected researcher of teacher effectiveness — who draws on his considerable background to explore the uses of value added data in high stakes decisions. Goldhaber provides a good, accessible description of the quirks in value-added models and shows how to improve the stability of the data, for example, by using multiple years of data. He also compares value-added measures to current methods for evaluating teachers, which he argues do little or nothing to distinguish good teaching from mediocre or just short of bad teaching, and finds value-added to be the better option.
My takeaway we should keep moving ahead to develop effective and fair means to link teachers with student performance. But we might want to ease off of the accelerator a bit.
Also check out the Center’s report, Promise or Peril: Teacher pay for performance plans. Patte Barth
