Challenges and validation in estimating teacher value-added in English secondary schools

Researchers in the United States have frequently used non-experimental data, including annual student test scores, to estimate teacher value-added (TVA). Experimental studies have largely validated this approach within the US context. Similar data are increasingly available in England. However, differences in the school systems, the way in which students are allocated to schools and teachers, and assessment practices present distinctive challenges for validly estimating TVA in England. This paper reviews the key issues that arise when applying TVA models to English secondary schools, including non-random student and teacher allocation and measurement error in assessments. In response, it proposes a practical framework of six empirical validation checks to assess the robustness of TVA estimation. These checks – focused on internal consistency, replication of empirical findings from elsewhere, and falsification tests – offer researchers a means to critically assess the plausibility of their estimates and improve the credibility of future TVA studies in the English context.