Data Handling and Modeling Issues

Because public schools do not randomly assign students and teachers across schools (methodological utopia), multilevel evaluation models which account for student and school contextual and practice variables in their natural settings provide the most rigorous means for empirically showing what is actually happening in school classrooms. Still, no statistical methodology can make up for faulty design or bad data. This article presents some important practical issues regarding data handling for multilevel analysis methodology. Also presented are important modeling design issues that need to be considered when applying hierarchical linear models (HLM) to the measurement of schools and for determining which factors impact the value schools add to students ’ achievement. he statistical method chosen for an analysis is usually a function of two things: the question being asked and the nature of the recorded data. In the case of measuring school effects, HLM is a multivariate regression-like analysis technique that was developed specifically for use in school effects research. HLM allows the examination of associations among multi-level, nested data such a