Skip to main content

The problem

Academic standards define what students are expected to know at each grade level, and they shape curriculum, instruction, and assessment. Edtech developers align their products to these standards, but a standards tag is often a claim, not a guarantee:
  • A standard is rarely a single skill
    • Most bundle several sub-skills (i.e., learning components)
    • A question could address one learning component and ignore the rest for a given standard
  • Standards vary by jurisdiction
    • A skill emphasized in one framework may be scoped differently, or absent entirely, in another
    • Alignment has to be judged against the framework in question, not a generic notion of a standard
Asking human experts to manually check educational content against every learning component, across a bank of hundreds of questions, does not scale.

What we’re building

Our Standards evaluators assess whether educational content aligns with a given standard’s learning components.
OutputDescription
Fine-grained, component-level dataValidates tagging and trust that an item measures the skill it claims to
Coverage mapsSurface exactly which standards, and which learning components, a question bank covers
Reasoning and feedback for every judgmentGives actionable feedback on turning looks aligned into is aligned
Our evaluators judge whether a question aligns to standards at the learning components level:
EvaluatorDescription
Math Judges whether a math question aligns to math standards in a supported jurisdiction
Literacy Coming soonJudges whether an English Language Arts question algins to ELA standards in a supported jurisdiction