Evaluating Teachers AND Administrators

Eric A. Hanushek
Published Date
Education Next
Recent school reform talk has focused importantly on teacher evaluations and on using evaluations for personnel decisions – both positive and negative. But this discussion is almost always too narrow. We should never focus exclusively on teacher evaluations without also including administrator evaluations. Parallel attention to administrator evaluations is essential for two reasons. First, it appears that principals have a very significant impact on student outcomes that cannot be ignored as we attempt to improve our schools. Second, it is important that administrators and teachers are working for the same outcomes, particularly when performance incentives are involved. While there is considerable anecdotal evidence that principals are important – including various movies about the charismatic principal or the bumbling bureaucrat, there has been very little systematic evidence about the magnitude of differences among principals or about their impact on student learning. Separating the influence of the principal from other factors including the choices of parents and teachers proves to be a very difficult task. In a recent article, Greg Branch, Steve Rivkin, and I use a variety of statistical approaches to isolate the impacts of a principal on student learning. The primary approach is to observe how student achievement in a school changes when the principal changes. In the most conservative estimation approach, a good principal (one in the top 16 percent of principal effectiveness) compared to an average principal annually gets the equivalent to an extra two months of learning out of students. Importantly, these gains hold for all of the students in school. While good teachers get similar gains compared to average teachers, those gains accrue just to the students in a single classroom. The other side also holds. An ineffective principal slows learning of students at the same rate. Thus, it is important to select and retain effective principals. The sheer magnitude of impact underscores the need for a reliable and accurate system for evaluating principals. As with teachers, it is difficult to predict principal performance by just looking at their background, training, and objective characteristics. Instead it is necessary to evaluate just how well the principal performs in the school – and to use these evaluations in pay and retention decisions. But, something left out of many policy discussions is the necessity of aligning the evaluation of principals with that of teachers. It would set up bad incentives to have the principal evaluating teachers without themselves being evaluated on student performance. If principals are evaluated primarily on other grounds than student achievement, they will not necessarily make decisions that are consistent with improving student performance. In this case, the common concern of teacher unions that decisions by principals can be arbitrary or can show favoritism in various ways may be justified. The simple conclusion: we need better evaluations of both teachers and principals, and we need to use these in personnel decisions. They have an enormous impact on the economic future of their students and of the nation. We need to reward and retain the best teachers and principals, and we need to move ineffective teachers and principals away from students. -Eric Hanushek