MDA
A controlled adaptive diagnostic that measures mathematical understanding at skill level.
MDA uses reviewed blueprints, MDA-exclusive published items, exact question versions, server-side evaluation, and evidence-aware item selection to build a defensible diagnostic picture without exposing answers or silently changing content mid-session.
Product status: these capability descriptions reflect the current development build and intended configured workflow. They are not a claim of measured customer results.
What MDA provides
Adaptive diagnostic engine
Difficulty, confidence, contradictory evidence, misconception probes, alternate representations, and information gain inform item selection.
Blueprint coverage
Readiness checks verify that required skills and item supply exist before a diagnostic is offered.
Exact-version delivery
Every response is evaluated against the immutable version that the learner actually saw.
Broad interaction support
Objective structured-response families share the same student renderer and server-safe evaluator registry.
Accommodation-aware selection
Item eligibility can account for authoritative learner accommodations without treating visual preferences as formal accommodations.
Operational health
Completion, abandonment, evaluator failures, exposure, selection reasons, coverage, and expired sessions can be monitored.
How it works
- 01
Validate diagnostic readiness
The blueprint, skill coverage, evaluator support, and published MDA-only pool are checked first.
- 02
Start or safely resume
A learner receives a server-owned session that expires and resets after the configured inactivity period.
- 03
Collect adaptive evidence
Each exact-version response updates the diagnostic state and informs the next eligible item.
- 04
Produce a student-safe result
The completed record summarizes skill evidence and appropriate next learning recommendations without revealing protected answer keys.
A clearer next step for every learner
- More precise instructional starting points
- Visible prerequisite gaps
- Safer adaptive measurement
- Actionable next-learning recommendations