Abstract:
The federal government's Race to the Top competition has promoted the adoption of test-based performance measures as a component of teacher evaluations throughout many states, but the validity of these measures has been controversial among researchers and widely contested by teachers' unions. A key concern is the extent to which nonrandom sorting of students to teachers may bias the results and lead to a misclassification of teachers as high or low performing. In light of this, it is important to assess the extent to which evidence of sorting can be found in the large administrative data sets used for VAM estimation. Using a large longitudinal data set from an anonymous state, we find evidence that a nontrivial amount of sorting exists - particularly sorting based on prior test scores - and that the extent of sorting varies considerably across schools, a fact obscured by the types of aggregate sorting indices developed in prior research. We also find that VAM estimation is sensitive to the presence of nonrandom sorting. There is less agreement across estimation approaches regarding a particular teacher's rank in the distribution of estimated effectiveness when schools engage in sorting.