Factors That Impact Teacher Effectiveness - 2221 Words

Jacqueline Adams Dickey 03/11/14 EDUC 661 Literature Review The increasing focus by many policymakers and measurement experts on using statistical models to evaluate teacher effectiveness has led to heightened debate regarding the usefulness of these methods but little discussion of the possible sources of error inherent in value-added modeling. Even the most robust models for making high-stakes decisions about teacher effectiveness contain numerous sources of error that can lead to ill-founded interpretations of the data. Both teachers and schools can be negatively impacted by the errant use of value-added modeling; increased awareness of possible limiting factors present in these models is needed, both at the policymaker and at the†¦show more content†¦As of 2013, 43 states measure teacher effectiveness, at least in part, through student achievement; 25 of these states are in the process of instituting entire teacher evaluation systems that rely on student growth data to measure teacher impact (National Council on Teacher Quality, 2013). Within the larger classification of student growth measures is value-added modeling (VAM), a method that uses a statistical model to establish a causal link between teachers and the achievement of students within their classroom. VAMs are considered promising because, in a perfect world, they might have the potential to promote education reform and to create a more equitable accountability system holding teachers and schools accountable for the aspects of student learning that are attributable to effective teaching without burdening teachers and schools with responsibility for factors outside of their control. Value-added models are being used to evaluate teacher effectiveness in districts across the country, including New York City and Washington, D.C. Opinion on their potential for misuse and misinterpretation has been split, with some arguing that VAMs are inherently imprecise and ill-suited to measuring teacher effectiveness (Rothstein, 2008; Baker et al., 2010), while others posit that the possibilities for error are minimal (Kane and