The Vocabulary evaluator gives developers fine-grained vocabulary insights that help ensure texts use words that align with grade-level expectations and support growth in academic language:
Estimates the background knowledge that a student at the target grade level is likely to have
Identifies complex words in the text (, , , etc.)
Evaluates overall vocabulary complexity relative to the background knowledge estimate
The evaluator was built and validated using the model and temperature below (other configurations will produce different results and may have lower accuracy):
List of Tier 2, Tier 3, archaic, and other complex words in the text
Complexity score
Slightly complex: Uses everyday, familiar vocabulary with few academic or domain-specific terms.
Almost entirely contemporary and conversational
Very low proportion of complex words (archaic, subject-specific, academic)
Easy to understand and does not impede comprehension of the bulk of the text (1-2 quick pauses for processing by the student may occur)
Moderately complex: Includes a mix of familiar and academic vocabulary, with some Tier 2 or Tier 3 terms that may require support.
Mostly contemporary and conversational
Low proportion of complex words (archaic, subject-specific, academic)
Generally allows students to comprehend the bulk of the text with little difficulty, though there may be occasional pauses for clarification (several quick pauses or occasional prolonged pauses may occur)
Very complex: Relies heavily on Tier 2 and Tier 3 vocabulary with limited contextual scaffolding.
Often unfamiliar, archaic, subject-specific, and/or overly academic
Often presents challenges that may slow down comprehension, but does not completely block the comprehension of the bulk of the text
Exceedingly complex: Uses dense academic and domain-specific vocabulary that is likely to be inaccessible without significant support.
This evaluator is provided as Early access. Comprehensive accuracy measures
are not yet available. Validation testing is ongoing.
We assessed performance against an expert-annotated dataset of 580+ texts from CLEAR Corpus ↗. Accuracy has been most extensively validated on Grades 3–4.
