Background: This paper develops a new method to estimate quasi-experimental evaluation models when it is necessary to control for measurement error in predictors and individual assignment to the treatment group is based on these same fallible variables. A major methodological finding of the study is that standard methods of estimating models that…

The relationships among factors reported on report cards developed for local school districts in Tennessee and how they related to student outcomes were studied. The study revealed few, if any, straight-line relationships between inputs and outputs reported. Report cards are vastly overrated as indicators of educational quality. (SLD)