Study Evaluates Advantages, Drawbacks of Value-Added Measures

A report released last month by the National Research Council and the National Academy of Education cautions policy makers against hasty adoption of value-added methods for evaluation of teachers, schools and students. In “Getting Value Out of Value-Added,” the product of the Workshop on Value-Added Methodology for Instructional Improvement, Program Evaluation, and Educational Accountability held in November 2008, researchers provide a comprehensive evaluation of the benefits and drawbacks of these evaluation systems. The report warns against using these models alone in “high-stakes” accountability assessments.

Value-added models have garnered considerable attention in recent years, first as a proposed alternative to the Adequate Yearly Progress (AYP) evaluations under No Child Left Behind and now as a possible answer to the Obama Administration’s call for use of student achievement data in the evaluation of teachers. Value-added assessments differ from “status” models, which simply take a snapshot of students’ achievement levels, and “growth” models, which compare these snapshots from year-to-year. Conversely, they not only track student achievement data over time, but also seek to isolate school, teacher and program effects and often take into account the impact of factors like poverty and family background.

According to workshop participants, value-added models share many flaws with their “status” and “growth” counterparts, including “the fact that tests are incomplete measures of student achievement,” but also encounter unique issues due to their statistical complexity. In particular, value-added estimates are often subject to large sampling errors, and may not be able to accurately assess teacher or school effectiveness. Students are not necessarily randomly distributed amongst teachers, for instance, and some classrooms experience high turnover. The report suggests that, in lieu of better assessments of student achievement and means of mitigating the aforementioned issues, policymakers use value-added models in “low-stakes” assessments of factors like professional development programs, but refrain from using them as the only measure for guiding “high-stakes” decision-making, like when to impose sanctions on a school or how to allocate teacher pay.