By Dale Chu
A new white paper on assessment innovation, Modernizing Measurement, argues that advances in technology and artificial intelligence have made traditional statewide summative tests increasingly obsolete. The paper—commissioned by Amira Learning and written by Whiteboard Advisors—calls for a shift toward “dynamic assessment” systems that are embedded into instruction, provide real-time feedback, and generate continuous data on student learning.
While the paper focuses more on policy and context than on tools (more in the final paragraph below), it cites the through-year assessment efforts of Maine, Delaware, Louisiana, and Montana among others as an exemplar, and is bullish about moving beyond a single summative exam. The idea seems to be to bundle the constellation of assessment experiments underway under a single modernizing label, as though placing the word “dynamic” in front of assessment resolves longstanding tradeoffs.
It does not.
Traditional statewide summative tests certainly have limitations. Results often arrive too late to shape classroom instruction, and the focus on reading and math has arguably led to a narrowing of the curriculum in many schools.
But the paper’s treatment of summative assessments borders on caricature. For starters, we are told that these tests are “primitive,” “instructionally inert,” and have “consumed billions of dollars.” Left unsaid is that statewide assessments typically cost states somewhere in the neighborhood of $25 to $35 per student annually—pennies on the dollars in the context of overall K-12 spending, and comparable to what districts already pay for a single high-quality ed-tech platform. The claim that testing is a major cost driver simply doesn’t hold up under even basic per pupil scrutiny.
This matters because many of the testing burdens educators experience today do not primarily stem from annual state tests themselves. They come instead from a sprawling ecosystem of diagnostics, benchmarks, interventions, and platform-based assessments that districts have layered on top of federal and state requirements. Ironically, many of those tools were developed by the same broader assessment and ed-tech ecosystem now arguing that testing has become too burdensome.
To be fair, some of the technological advances described in the paper are real. Adaptive testing, AI-supported scoring, and more sophisticated growth measures may well improve parts of the assessment landscape. States should absolutely explore ways to make measurement more meaningful. But innovation requires clarity, not buzzwords. Right now, dynamic assessment feels less like a clearly defined category and more like an aspirational branding exercise.
The paper closes by teasing a forthcoming Part II that promises a deeper look at what dynamic assessment looks like in practice. That’s welcome news. Because before states embrace dynamic assessment, it would be helpful to know exactly what it is.