Value-Added Models and Accountability: Next Steps

Much basic research remains to be done to build high-quality value-added models of the productivity of schools, teachers, and administrators. This project is organized around three major interrelated research activities: (a) Theory and development of value-added models; (b) Testing, demonstration, collaboration, and implementation of innovative value-added models with 4 key research partners: the Chicago Public Schools, Dallas Independent School District, Milwaukee Public Schools, and Minneapolis Public Schools; and (c) Dissemination of research findings and applications.

The value-added models we are developing have been identified as high priority areas of research by value-added researchers and by the educators and stakeholders that we have worked with over the past two decades. One of our most important research priorities is to develop and refine a new value-added model that combines the best features of models with random and fixed individual effects—that is, high precision and low selection bias. We refer to the new model as the generalized value-added model with conditional random effects and multivariate shrinkage (CRE-MS). We expect that it will yield important reductions in expected mean squared error relative to other value-added models, including those based on fixed or random effects.

One further advantage of the model is that it provides estimates of precision that take into account the error due to uncontrolled selection bias as well as traditional estimation error. In addition, we propose to extend the value-added framework to accommodate: (a) mid-year testing; (b) multidimensionality in measured achievement, possible violations of the equal interval property, and irregularities in horizontal and vertical equating; and (c) multiple assessment systems in the same district, including short-cycle assessments.