Chapter 6: Value-Added Models and the Next Generation of Assessments

This chapter discusses some of the fundamental features of value-added models, with particular focus on the interaction between the design and interpretation of value-added models and the design and properties of student assessments. We present a case study using actual state and district data from Wisconsin.

A value-added model is a quasi-experimental statistical model that yields estimates of the contribution of schools, classrooms, teachers, or other educational units to student achievement (or other student outcomes), controlling for other (nonschool) sources of student achievement growth, including prior student achievement and student and family characteristics. The model produces estimates of school productivity—value-added indicators—under the counterfactual assumption that all schools serve the same group of students. This facilitates apples-and-apples school comparisons rather than apples-and-oranges comparisons. The objective is to facilitate valid and fair comparisons of student outcomes across schools, given that the schools may serve very different student populations.