Implementing value-added measures of school effectiveness: getting the incentives right

As part of their efforts to hold schools accountable, several states now calculate and publicize value-added measures of school effectiveness. This paper provides a careful evaluation of the value-added approach to measuring school success with particular attention to its implementation as a tool for increasing student achievement. In practice, even the more sophisticated of the measures currently in use fail to account for differences in resources, broadly defined, across schools and to address the problem of measurement error. The authors find that, as implemented, value-added measures of school effectiveness distort incentives and are likely to discourage good teachers and administrators from working in schools serving concentrations of disadvantaged students. The authors use a large longitudinally-matched data set of fifth grade students in North Carolina to document that approximately two-fifths of the differentially favorable outcome for schools serving advantaged students result from statistical bias associated with measurement error and that correcting for the measurement error leads to significant changes in the relative rankings of schools. (C) 2001 Elsevier Science Ltd. All rights reserved.