Exploring Models of School Performance: From Theory to Practice. CSE Report 673.

The findings and opinions expressed in this report are those of the authors and do not necessarily reflect the positions or policies of the Our purpose in this report is to present and discuss competing accountability approaches, or models, designed to systematically indicate how a school's students are performing academically. Within the framework of the current federally mandated accountability legislation, increased interest in models measuring school performance has caused educational policymakers to consider several key issues. These issues include whether results from different accountability models yield different inferences about a school's performance; what assumptions underlie each of the models; how different models are implemented; and ultimately which model is best suited for a particular context. We address these issues by building a framework for accountability models and then explicitly comparing and contrasting these competing models. In order to accomplish this, we first need to examine two distinct pieces of the larger puzzle. With the first piece, we briefly summarize previous research on school performance. This is done in order to ground all of the accountability models and provide some reference for considering how an accountability model might be constructed. With the second piece, we present building blocks for accountability models. These building blocks include a) important properties of assessments, b) test metrics, c) ways of summarizing student achievement, and d) monitoring achievement growth over time; all of which need to be considered before they are incorporated into an accountability model. Once we have the foundation and building blocks in place we can examine the continuum of accountability models, each of which results in a performance indicator. We consider the choice of model as lying on a continuum because accountability models range from simple calculations on the one end to complex statistical models on the other. At the upper end of the spectrum is a set of accountability models known as value-added models (VAM), which we compare separately. We also compare inferences based on one of these VAMs against inferences based on current federally mandated accountability models. 1 Examining competing accountability models and linking them back to the foundations and building blocks leads to both theoretical and practical implications that are central in considering which model is most appropriate for a given (physical and political) context. One fundamental concern is whether the accountability model can accurately capture the academic progress of underprivileged students (e.g., low socioeconomic status [SES]) and, by extension, …