Using fuzzy logic to identify schools which may be misclassified by the No Child Left Behind Adequate Yearly Progress policy.

摘要

This investigation developed, tested, and prototyped a Fuzzy Inference System (FIS) that would assist decision makers in identifying schools that may have been misclassified by existing Adequate Yearly Progress (AYP) methods. This prototype was then used to evaluate Louisiana elementary schools using published school data for Academic Year 2004.;Many authorities believe that the increasing possibility of schools being misclassified with respect to adequate yearly progress (AYP) is a major weakness of the No Child Left Behind (NCLB) act. Human judgment is being introduced into many AYP methodologies. The difficulty of aggregating human judgment with high stakes testing results is widely known (Ammar, 2004). This investigation used a modified Analytic Hierarchic Process (AHP) to capture the human judgments and then trained Fuzzy Inference System modules using neural networks to match that judgment. The modules in turn were asked to evaluate the 2004 Louisiana School Performance data and identify schools that may have been misclassified.