False Performance Gains: A Critique of Successive Cohort Indicators

There are many ways to use student test scores to evaluate schools. This paper defines and examines different estimators, including regression-based value-added indicators, average gains, and successive cohort differences in achievement levels. Given that regression-based indicators are theoretically preferred but not always feasible, we consider whether simpler alternatives provide acceptable approximations. We argue that average gain indicators potentially can provide useful information, but differences across successive cohorts, such as grade trends, which are commonly cited in the popular press and used in the Safe Harbor provision of federal school accountability laws, are flawed and can be misleading when used for school accountability or program evaluation