How CheckReadability Computes Consensus Grade Without Scale Distortion

Every single readability formula has known edge-case blind spots: ARI penalizes 8-letter 1-syllable words ('strengths'), SMOG jumps on short 3-syllable words ('idea'), and FORCAST ignores sentence length. CheckReadability's Consensus Grade averages the seven formulas whose native output is a direct US school grade level—FKGL, Gunning Fog, SMOG, Coleman-Liau, ARI, Linsear Write, and FORCAST—canceling out single-formula variance while strictly isolating non-grade index scales (FRE 0–100, Dale-Chall raw index, and LIX) so incompatible units never corrupt your grade.
Key Compliance & Engineering Takeaways
- Unit Purity Rule: Naive scripts that average `FRE = 65.0` or `LIX = 38.0` into US school grades (`~8.0`) produce garbage numbers—only dimensional school-grade outputs belong in a grade mean.
- Balanced Feature Ensemble: Combines 2 alphanumeric character formulas (ARI, Coleman-Liau), 2 total-syllable/point formulas (FKGL, Linsear Write), 2 polysyllable formulas (Gunning Fog, SMOG), and 1 monosyllable vocabulary formula (FORCAST).
- Specialist Formula Preservation: Spache (calibrated strictly for Grades 1–3) is displayed in its own dedicated card rather than dragging adult Grade 14 legal briefs downward.
- Direct CEFR A1–C2 & Reader Age Mapping: Maps Consensus Grade transparently onto international CEFR language proficiency bands (`<3.5 A1`, `<5.5 A2`, `<8.0 B1`, `<11.0 B2`, `<14.5 C1`, `14.5+ C2`).
Mathematical & Algorithmic Specification
Exact peer-reviewed equations & linguistic rules executed locally inside CheckReadability's browser engine.
CheckReadability 7-Formula Consensus Grade
Verified SpecConsensus = ( FKGL + GunningFog + SMOG + ColemanLiau + ARI + LinsearWrite + FORCAST ) / 7Variables: Strictly averages the 7 formulas whose native regression output unit is US School Grade Years (0–18+)
Psychometric & Engine Note: Excludes Flesch Reading Ease (0–100 inverse ease), New Dale-Chall (4.9–10+ Cloze raw score), LIX (20–60 index), Fry (2D zone), and Spache (K–3 primary specialist) from raw arithmetic mean.
All 12 CheckReadability Formulas: Output Scale & Consensus Inclusion
| Formula Name | Native Output Scale | In 7-Formula Consensus? | Primary Linguistic Signal |
|---|---|---|---|
| 1. Flesch-Kincaid Grade (FKGL) | US School Grade | Yes (Included) | Words/sentence + syllables/word (75% comprehension) |
| 2. Gunning Fog Index | Years of Education | Yes (Included) | Words/sentence + % 3+ syllable complex words |
| 3. SMOG Readability Index | US School Grade | Yes (Included) | Square-root 3+ syllable density per 30 sentences (100% comp.) |
| 4. Coleman-Liau Index (CLI) | US School Grade | Yes (Included) | Letters per 100 words + sentences per 100 words |
| 5. Automated Readability (ARI) | US School Grade | Yes (Included) | Alphanumeric chars/word + words/sentence |
| 6. Linsear Write Formula | US School Grade | Yes (Included) | 1-pt easy words vs. 3-pt hard words per sentence |
| 7. FORCAST Formula | US School Grade | Yes (Included) | 1-syllable word ratio (punctuation-independent) |
| 8. Flesch Reading Ease (FRE) | 0–100 Inverse Index | Displayed Separately | Higher = easier; averaging 65 into Grade 8 would break math |
| 9. New Dale-Chall | 4.9–10+ Cloze Index | Mapped via Lookup Band | % unfamiliar words outside 3,000-word 4th-grade list |
| 10. Fry Readability Graph | 2D Coordinate Zone | Plotted Visually (SVG) | (Sentences/100w, Syllables/100w) curve + invalid-zone check |
| 11. LIX & RIX | 20–60 Index / Ratio | Displayed Separately | Multilingual >6-character word index |
| 12. Spache Primary Formula | Grades 1.0–3.9 Scale | Displayed Separately | Early-elementary K–3 specialist sight-word regression |
Why Ensemble Averaging Outperforms Any Single Readability Formula
In machine learning and psychometrics, combining multiple uncorrelated or weakly correlated estimators reduces variance and prevents adversarial gaming.
Suppose a writer splits every 30-word sentence into 6-word fragments (`'Pharmacogenetic contraindications necessitate oversight. Indemnification provisions apply.'`). Sentence length drops to 4.5 words—tricking sentence-heavy formulas—yet FORCAST, SMOG, and Coleman-Liau immediately catch the 5-syllable / 15-character vocabulary load. Conversely, if a writer strings twenty 1-syllable words together with 'and... and... and...', FORCAST stays low while FKGL, Fog, and ARI flag the syntactic sprawl. Averaging all seven grade-scale estimators produces a balanced, manipulation-resistant reading grade.
Run These Offline Readability & Compliance Tools Now
100% browser-local execution with built-in sample documents — 0 bytes uploaded
CEFR (A1–C2) & ESL Reading Level Estimator
Maps English text complexity onto CEFR A1, A2, B1, B2, C1 & C2 bands for learners & teachers.
Side-by-Side Draft Diff & Readability Score Comparator
Compare Original vs. Edited drafts with word-level redlines and 12-formula score deltas.
Frequently Asked Questions
Why isn't Spache included in the general Consensus Grade?
George Spache's regression (`0.121 × ASL + 0.082 × Unfamiliar% + 0.659`) has small slope coefficients calibrated strictly on Grades 1–3 children's books; even on a college legal contract, Spache outputs ~Grade 5–7, which would artificially pull adult consensus grades down.
Is the Consensus Grade included in PDF and CSV exports?
Yes—both the client-side PDF Audit Report and Batch CSV Scorecard report the Consensus Grade, Grade Label, Age Range, and CEFR band alongside individual formula scores.
Does `POST /api/score` return the exact same Consensus Grade as the browser UI?
Yes—both the browser Studio and `/api/score` import the exact same `analyzeReadability()` function in `src/lib/engine/analyzer.ts`.
More Formula & Linguistic Engineering Specifications

SMOG vs. Flesch-Kincaid: Why 100% vs. 75% Comprehension Shifts Scores by ~2 Grades
Mathematical and psychometric breakdown of why Harry McLaughlin's SMOG Index consistently reports 1.5 to 2.5 grade levels higher than Flesch-Kincaid Grade Level on identical text.

Why Readability Calculators Disagree: Silent-E, Affixes & Syllable Counting Algorithms
Deep dive into English orthographic-to-phonemic syllable counting: how silent '-e', '-ed'/'-es' inflections, vowel hiatus ('create', 'chaos'), syllabic '-le', and exception dictionaries govern Flesch, Fog, and SMOG accuracy.

Character Formulas (ARI, Coleman-Liau, LIX/RIX) vs. Syllable Formulas Benchmarked
Mathematical and linguistic comparison of character-length formulas (ARI, Coleman-Liau, LIX, RIX) against syllable-based formulas (Flesch-Kincaid, SMOG, Gunning Fog).

Scoring UI Microcopy, Forms & Bullet Lists Without Full Sentences (FORCAST)
Mathematical specification of Caylor et al.'s (1973) FORCAST formula: why sentence-based formulas fail on UI strings, survey items, and slide decks, and how monosyllable ratios solve UX readability.

Fry Readability Graph vs. Lexicon Formulas (New Dale-Chall & Spache) Explained
Technical specification of Edward Fry's 100-word visual coordinate graph (including invalid zones), Chall & Dale's 3,000-word +3.6365 step function, and George Spache's unique sight-word formula.

Computational Style Diagnostics: Passive Voice POS Tagging & Lexical Density (Ure / TTR)
How CheckReadability combines Compromise Part-of-Speech (POS) tagging to eliminate false-positive passive voice ('was tired', 'is closed') alongside Ure Lexical Density and Type-Token Ratio.