As a general rule, conservatives who work on K-12 education tend to embrace high-stakes tests that are designed to assess what students have and have not learned over the course of a school year. When critics warn of the danger that teachers will “teach to the test,” proponents of high-stakes will generally reply that, well, if the test is designed to assess whether students have mastered important skills and bodies of knowledge, that’s not obviously a bad thing. This consensus around high-stakes testing has contributed to enthusiasm for value-added assessment, in which teachers are assessed in part on the basis of how their students fare on standardized tests. Marcus Winters, an assistant professor in the College of Education at the University of Colorado Colorado Springs and a senior fellow at the Manhattan Institute, is one of the leading proponents of this approach.
I am increasingly convinced that while high-stakes testing may have once had an important role to play, it has outlived its usefulness. Rick Hess, Michael Petrilli, and Martin West, among others, have warned that mandating that teachers be evaluated on the basis of value-added analysis might stifle innovation, particularly as it relates to how schools deploy personnel. Value-added analysis built around the model of a single teacher leading a class of 25 students might make it impossible for schools to pursue more personalized instructional models, or models in which a team of professionals is responsible for raising achievement. It is easy to see why value-added analysis might represent an improvement over the status quo if we assume a fixed instructional model. But if we instead believe that we need to find entirely new ways to do things, and to let a thousand flowers bloom through charter schools and course-level instructional choice, the danger involved in locking ourselves into a fixed model becomes clear.
The long term vision is software-based assessments that are “Embedded, low-anxiety, formative – not the assessment event that creates stress for teachers, principals, students, and families,” said Cameron Evans of Microsoft, and that serve as “durably predictive short term metrics,” said John Katzman of “real-world outcomes” like happiness, employability, and civic responsibility. “We have to think about assessment as something that should bolster education, not just measure it.”
That is, instead of high-stakes tests administered at long intervals, Anya argues that we ought to move towards low-stakes software-based assessments that don’t need to be administered at all. Rather, they will be part and parcel of the work students do in the classroom. Anya’s vision of software-based assessment is closely aligned with the advent of the “flipped classroom,” a model in students view short video lectures at home while devoting their in-class time to completing exercises and projects, and interacting with teachers and fellow students.
High-stakes testing sounds tough, and I absolutely get why many education reformers gravitate towards tools and institutions that appear to toughen up the often rather soft world of K-12 education. But the explosion of computing power means that we can collect more data more often, and this enables us to a do a better, finer-grained job of understanding what students actually know. It is important to note that while opposition to high-stakes testing and value-added analysis often seems self-serving — it is easy to see why ineffective teachers might resist accountability — moving towards embedded software-based assessment actually raises the level of transparency, by allowing us to monitor not just what happens on the day of a high-stakes test, but rather to see how students learn over time. While high-stakes testing lends itself to fraud, software-based assessment makes it, in theory at least, extremely difficult.