Examining the Added Value of Value-Added Models
For decades parents, educators, policymakers, taxpayers, and researchers have wrestled with how to fairly evaluate teachers. Attempting to judge teachers for the academic development of a student, and not just absolute achievement, state departments of education have begun using a statistical technique known as value-added modeling, or VAM). The increasing implementation of VAM to evaluate teachers has touched off a maelstrom of debate.
Arguments surrounding VAM have typically focused on the philosophical, statistical, and procedural merits of VAM, but overlooked a fundamental question: How can principals, the individuals responsible for selecting, developing, and evaluating teachers, actually utilize VAM?
The answer seems straightforward. Oft-cited examples from the world of professional sports illustrate the potential of dispassionate algorithms to strip away the limits of observation, biases of reputation, and influences of circumstance to fairly evaluate performance. Proponents of VAM suggest that principals could apply these evaluation precepts to improve the teaching corps of a school.
To test the utility of VAM ratings in the process of principals evaluating current or prospective teachers, suspend any disbelief in VAM. Assume that VAM scores accurately reflect the ability of teachers to foster academic growth in their students. If VAM provided valid and fair evaluations of teacher performance, then principals could identify and replace teachers with unsatisfactory VAM ratings.
However, to justify teacher replacements, principals need to know that the VAM scores of the incoming teachers exceed those of the incumbents. Without this additional information, replacing a teacher relies on principals' observations and judgments, which critics have characterized as arbitrary and capricious.
As an example, consider the options confronting a high school principal evaluating an algebra teacher who earned an unsatisfactory VAM rating. The principal could swap the algebra teacher with another teacher from within the building. However, if the replacement teacher has not taught algebra, then he or she lacks a VAM rating. This replacement teacher could perform the same as, or even worse than, the original teacher.
Even if this principal attempted to terminate the ineffective teacher, a similar challenge results. Rookie teachers lack a comparable VAM rating. Without VAM results for the applicants, improved algebra teaching relies on the capabilities of the principal to choose an effective candidate from the applicants. In reality, this method of teacher selection was used to hire the original, ineffective algebra teacher.
Furthermore, what if that ineffective algebra teacher earned a better VAM rating than anyone who applied to the position? Even worse, across the country available teaching positions remain vacant. When principals cannot fully staff their schools, available teachers become invaluable, regardless of their VAM ratings.
Advocates have oversold the utility of VAM to help improve a school's teaching corps. In our conversations with principals, they have appreciated the ability of VAM to highlight the effectiveness of the overlooked or unpopular teacher, as well as expose the well-liked, but inept. Yet, when principals decide who will teach our children, VAM offer them little to no value.
Craig Hochbein is an assistant professor of educational leadership at Lehigh University and an assistant editor of the international journal School Effectiveness and School Improvement. Abby Mahone is currently pursuing a doctoral degree in educational leadership at Lehigh University. She was a classroom teacher for 10 years.