Consulting
Continuous Measurement partners with ed-tech, curriculum providers, assessment companies, and policymakers to build or improve learning and measurement systems that leverage the best of what cognitive science, psychometrics, and AI/ML have to offer.
Start a ConversationThe problem we solve
The education sector is awash in tools, data, and AI — yet coherent understanding of what works, and for whom, remains as elusive as ever.
What We Do
Most learning systems are built from the outside in — tools selected, data collected, and measurement questions asked afterward. Continuous Measurement works the other way. We start with the construct: what is being learned, how it develops, where it typically breaks down, and what evidence would actually tell you whether it's been acquired.
That foundation — the content and measurement architecture — is what most ed-tech projects skip, and what makes the difference between a system that produces actionable data and one that produces noise. We do three things, in any combination a partner needs:
Content and measurement architecting — defining the cognitive architecture of a learning domain: the construct map, the prerequisite structure, the misconception taxonomy, the difficulty parameters, and the data schema that downstream tools need to work from. This is the work that makes assessment defensible and ML models interpretable.
System review and alignment — auditing existing tools, instruments, and data pipelines against a defined measurement framework. We identify where the system is incoherent, where data structures are incompatible, and where the connective tissue is missing — and we produce a prioritized roadmap for making it work.
Measurement design and build — designing and developing instruments, pipelines, and data architectures from specification, including the labeled data schemas and latent state definitions that ML models require. We build for calibration and downstream use, not just for scoring.
Service Lines
Our engagements draw from three interlocking capabilities — architecting the content and measurement foundation, reviewing and aligning what exists, and designing or building the instruments and data infrastructure that make systems work.
Work Mode 01
Defining the cognitive and measurement foundation that everything downstream depends on — before tools are selected, data is collected, or models are trained.
Defining what is being learned, how it develops, where it typically breaks down, and what evidence of acquisition actually looks like. Deliverables include construct maps, prerequisite structures, cognitive step sequences, and misconception taxonomies — the specification layer that makes assessment defensible and ML models interpretable.
Designing the labeled data structures and latent state definitions that adaptive and predictive systems need — before the pipeline is built. We specify what each response represents, how states are defined, and what the model is actually being trained to predict, so that downstream ML is interpretable and defensible to assessment-literate buyers.
Specifying the difficulty continuum for a learning domain — the parameters that govern how construct complexity scales across items, tasks, and grade levels. This is the scaffolding for IRT calibration, adaptive sequencing, and growth modeling.
Work Modes 02 & 03
Auditing what exists and aligning it to a measurement framework — or designing and building the instruments, systems, and pipelines that don't exist yet.
Auditing existing tools, instruments, and data pipelines against a defined measurement framework. We identify incoherence, incompatible data structures, and missing connective tissue — and produce a prioritized roadmap for making the system work. Includes assessment ecosystem audits, data flow mapping, and integration specification.
Designing and building diagnostic instruments from specification — with full cognitive architecture documentation, construct-level distractor mapping, IRT parameter priors, and response pattern interpretation guides. Built to be calibrated, not just scored. Spans formative checkpoints through longitudinal growth instruments.
Building AI and ML products, adaptive learning systems, and diagnostic pipelines where the measurement foundation is in place before the first line of code is written. Includes knowledge tracing models, recommendation architecture, and the labeled data scaffolding that makes model outputs interpretable and defensible.
In Practice
These are representative examples of the problem types we're built for — not case studies, but illustrations of how the two modes play out.
Architecting & Build
Diagnostic Data Architecture for an Adaptive Learning Platform
Review & Align
Connecting Curriculum, Assessment, and Intervention Across a District System
Design & Build
Psychometric Specification for a New Diagnostic Instrument Suite
Review & Advising
Portfolio Coherence Review for an Education Funder
Who We Work With
Our partners share a common challenge: they need measurement infrastructure that's defensible, data that's actionable, and systems that hold together under scrutiny.
Technology
Building adaptive systems, diagnostic tools, or AI-powered learning products that need defensible measurement architecture and labeled data structures from first principles.
Curriculum
Developing instructional materials that need embedded diagnostic instruments built to psychometric standards — not intuition-based exit tickets.
Assessment
Modernizing item banks, redesigning data architectures, and building the interpretive frameworks that make results legible to the practitioners who need to act on them.
Policy & Philanthropy
Evaluating measurement coherence across education investments, developing theory-of-change frameworks, and assessing whether existing tools are producing data worth acting on.
Let's Work Together
We don't propose solutions before we understand the architecture. Tell us about your context — the tools you're using, the data you're generating, and the questions you can't currently answer — and we'll tell you honestly whether we can help.