Readability Formulas and How to Read Each Score

Readability Formulas and How to Read Each Score

A readability score looks like a single verdict, but each formula answers a slightly different question with different inputs. Flesch-Kincaid Grade Level and Flesch Reading Ease both come from average sentence length and average syllables per word, one expressed as a US school grade and the other on a 0 to 100 scale where higher means easier. SMOG counts only words of three or more syllables across a sample of sentences, which is why health writers rely on it, and Coleman-Liau and the Automated Readability Index count characters instead of syllables.

That is also why the same draft can score grade 8 on one formula and grade 11 on another without either being wrong. The readability scorer reports Flesch-Kincaid Grade Level, Flesch Reading Ease and SMOG; the sections below explain how each of those is calculated and what range to aim for, then cover the other formulas you will meet in style guides and compliance checklists.

Check Your Flesch-Kincaid Grade Level

Run your text through the readability tool above and it returns a Flesch-Kincaid Grade Level score instantly, the estimated US school grade needed to understand what you wrote. A score of 8 means an average 13-year-old can read it, and most general web content targets grades 6 to 10 while academic papers may reach grade 12 or higher.¹

How the Flesch-Kincaid formula translates sentence structure into a grade number

The Flesch-Kincaid Grade Level formula calculates its output from two measurable properties: average sentence length and average syllables per word. Its exact form is: 0.39 × (total words ÷ total sentences) + 11.8 × (total syllables ÷ total words) − 15.59. A text where every sentence averages 20 words and every word averages 1.5 syllables scores approximately 8.5, placing it at a standard web content reading level. Both coefficients matter, but shortening sentences produces larger score reductions per edit than switching individual multisyllabic words, because average sentence length responds directly to each split or deletion.

The constant −15.59 calibrates the formula against comprehension studies conducted on American schoolchildren in the 1970s. Kincaid and his colleagues at the US Navy tested reading comprehension across grade levels and adjusted the constants until formula outputs matched observed reading ability at each grade. Because the formula was calibrated on English text and American school grades, results are most reliable for native English writing and least reliable for translated content or text dense with proper nouns, which are often multisyllabic but rarely impede comprehension for familiar names.¹

Why sentence length drives grade level more than word complexity

Short words alone do not produce a low grade level. A paragraph built entirely from one-syllable words but structured as 60-word run-on sentences still scores at grade 12 or above, because average sentence length dominates the formula for most practical writing. Conversely, a paragraph of 15-word sentences that includes occasional technical terms stays below grade 8 even when several words push the syllable average up. This asymmetry tells you where to focus first when rewriting: cut sentence length before you tackle word choice.²

Target grade levels for different content types and audiences

Matching your grade level target to your audience matters more than chasing a single universal number. General-interest journalism targets grade 6–8; major online news outlets cluster around grade 8 for main news articles. Consumer product pages and marketing copy perform best at grade 5–7, where friction is lowest for impulse decisions. Developer documentation tends to sit at grade 8–10 because technical terms are unavoidable, though well-written API reference docs with clear sentence structure stay at the lower end of that range.³

Medical and legal text scores high by necessity, not by poor writing. A drug interaction warning must include the chemical name even if it adds syllables, and a contract clause must preserve precise meaning even if that requires a 35-word sentence. For these content types, targeting grade 12–16 is realistic and appropriate. If you write supplementary patient-facing content alongside clinical documentation, keep that explanatory layer at grade 6–8 while leaving the clinical core untouched.

Grade level benchmarks from published style guides

The US federal plain language guidelines recommend grade 8 or below for all government public communications. For news writing, the American Press Association style guide targets grade 8 for general stories and grade 5–6 for broadcast scripts intended for listening rather than reading. Microsoft's writing style guide targets grade 7–8 for user-facing documentation. These benchmarks come from reader comprehension research, not from SEO goals, which gives them more durability as targets than grades derived from search performance data alone.⁴

Why Flesch-Kincaid diverges from perceived reading difficulty, and how to supplement it

Flesch-Kincaid measures two surface-level text properties, not the actual difficulty of your ideas. A passage explaining quantum entanglement in short sentences and common words scores at grade 5 while remaining genuinely hard to understand for most adults. A passage describing a simple recipe with occasional French cooking terms scores grade 9 while being perfectly accessible to anyone who cooks. This gap between measured grade level and actual comprehension difficulty is a known limitation of all syllable-and-sentence-length formulas.

The formula treats all syllables equally. "Photosynthesis" and "entertainment" are both five syllables and both push the score up, but entertainment is familiar to virtually all English speakers while photosynthesis requires domain knowledge to interpret correctly. Similarly, highly specific technical abbreviations such as "IP," "API," or "HVAC" contain few syllables but present real comprehension barriers for readers outside the domain. You should use Flesch-Kincaid as a starting signal, not as the sole readability verdict.

Combining Flesch-Kincaid with engagement data and real-world calibration

The most reliable readability check combines a formula score with actual engagement data from your analytics platform. Pages scoring at grade 8 with low bounce rates and strong time-on-page metrics confirm that grade 8 works for your specific audience. If grade 6 pages underperform on engagement, the problem likely lies in content structure, argument quality, or page layout rather than reading difficulty. Flesch-Kincaid describes the text; your analytics tell you whether the text works for the people reading it.

For content that must comply with accessibility standards, WCAG 2.1's Reading Level criterion recommends providing supplementary content when text requires more than a lower secondary education level to understand. To keep accessibility visible, measure your Flesch-Kincaid grade as a standing step in your content production checklist rather than treating it as an afterthought. A high grade level is acceptable when your entire audience has the domain knowledge to read it, such as a medical journal targeting physicians or a legal service for practicing attorneys. The mismatch occurs when specialized language targets a general audience.⁵

When to use this

Run your draft through the tool above whenever you need to interpret a readability score, confirm what grade level your writing actually targets, or adjust your content until it matches your intended audience.

Examples

You receive scores: 5.2, 8.7, 12.4, 15.1. What do they mean?

Grade 5.2: Suitable for 11-year-olds (very easy). Grade 8.7: Standard web content (14-year-old reading level). Grade 12.4: Academic and legal text. Grade 15.1: Specialized academic content that may require graduate-level education. Most popular blogs and news sites target grades 6–8.

What grade level should you aim for?

General public content: Grade 6–8. Technical documentation: Grade 8–10. Academic papers: Grade 12+. Legal/medical texts: Grade 14+ (unavoidable for specialized terminology). Marketing copy: Grade 5–7 for broadest reach.

Compare Your Reading Ease and Grade Level Scores

Run your text through the tool above and you get both scores at once: Reading Ease on a 0 to 100 scale, and Grade Level expressed as a US school grade. Both draw on the same two inputs, sentence length and word complexity, but present the result differently, and reading them side by side tells you more than either number alone.¹

The shared formula inputs behind both scores

Both the Flesch Reading Ease and the Flesch-Kincaid Grade Level use exactly two text measurements: average sentence length in words and average word length in syllables. Reading Ease was published in 1948 by Rudolf Flesch; Grade Level was derived from it in 1975 by Kincaid for the US Navy to help standardize technical manual readability. The Reading Ease score runs from 0 to 100, where higher numbers indicate easier text, while Grade Level produces a positive number representing the US school grade needed to understand the document. Mathematically, these two scores are strongly inversely correlated: as Grade Level rises, Reading Ease falls.¹

Understanding the coefficient differences explains the occasional discrepancy between them. Grade Level weights sentence length more heavily relative to syllable density, while Reading Ease weights syllable density more heavily. Long-sentence simple-word texts therefore produce starkly different values: a bureaucratic document with 40-word sentences of mostly single-syllable words scores high in Grade Level but moderate in Reading Ease. Recognizing this pattern tells you which editing technique to apply first.

When the two scores seem to contradict each other

A text can score 60 in Reading Ease (standard difficulty) and grade 12 in Grade Level (college level) if it uses many long sentences composed primarily of short words. This pattern is common in bureaucratic and legal writing, where sentence structure is complex but vocabulary is relatively accessible. When you see this combination, sentence splitting brings both scores into alignment more effectively than vocabulary substitution. The inverse pattern, low Reading Ease with a moderate Grade Level, points to dense vocabulary in shorter sentences, which is common in scientific abstracts.

The same logic helps you decide what not to change. Writers often reach for a thesaurus when a score looks bad, but swapping a precise term for a simpler one can damage accuracy without moving the numbers much. If the document has long sentences and accessible words, the improvement lives in structure, not vocabulary, and that is the cheaper edit to make.

Reading Ease score ranges and what each means in practice

Reading Ease divides into five practical ranges. Scores from 90 to 100 indicate very easy text, roughly equivalent to 5th-grade material and suitable for young readers or simple instructions. The 70 to 90 range covers easy text appropriate for general consumer content and casual blog writing. Scores from 60 to 70 are the standard zone for most general-audience web content, mainstream journalism, and product descriptions for non-specialist buyers. Below 60, text grows progressively more difficult: the 30 to 60 range covers academic writing and policy documents, while scores below 30 require specialized knowledge.

For most web content, the 60–70 zone is your practical target. Falling into the 50s does not mean your content is bad, but some readers in your audience will work harder to extract the information they need. Pages where readers must work harder show this in their analytics: higher early-exit rates on mobile, shorter session durations, and lower return visit rates. Reading Ease gives you a number to optimize before publishing rather than discovering readability problems through post-launch data.²

Reading Ease benchmarks for popular publication types

Reader's Digest has traditionally targeted reading ease scores around 65 to match its broad adult audience. Time Magazine typically scores about 52, while the Harvard Law Review sits in the low 30s. Scientific American ranges from 45 to 55 for its accessible science articles. For your own content, benchmark against publications your target audience already reads comfortably: if they subscribe to mid-complexity trade publications, a score in the 50–65 range matches their habitual reading experience.

Choosing between Reading Ease and Grade Level, and using both together

Reading Ease is the better metric for general audiences because its 0–100 scale maps naturally to percentage-style thinking: a score of 65 feels like a passing grade, which is easy to explain to non-technical stakeholders and editorial managers. Grade Level is more useful when you need to match content to a specific education level, such as writing to an 8th-grade health literacy standard or calibrating employee communications for a workforce with a known education profile. For regulatory and compliance contexts, Grade Level appears more frequently in formal specifications, including the SEC's Plain English Handbook, which sets out plain English principles for investor disclosures without specifying a numerical grade level.³ Federal plain-writing rules follow the same pattern: the Plain Writing Act of 2010 requires covered documents to be "clear, concise, well-organized," and it deliberately defines plain writing in qualitative terms rather than mandating any readability score⁴.

Running both scores on the same text lets you compare Reading Ease and Flesch-Kincaid and identify which type of complexity is driving readability down. High Grade Level combined with reasonably high Reading Ease (above 50) points to long sentences as the primary problem: splitting those sentences raises Reading Ease and lowers Grade Level simultaneously. Low Reading Ease combined with a moderate Grade Level indicates that vocabulary density is the issue rather than sentence structure. A third pattern, very low Reading Ease and very high Grade Level in text that does not feel academic, often signals embedded lists or formatted elements that fragment sentence count without reducing actual density.

Prioritizing which score to optimize first

If both scores are outside your target range, address sentence length first. Sentence length changes immediately improve both Grade Level (lower) and Reading Ease (higher), and they are editorially safer than vocabulary changes because splitting a sentence almost never changes its meaning. Several US states have statutes specifying minimum Flesch Reading Ease scores for certain document types, such as Florida Statute 627.4145, which requires a minimum score of 45 for insurance policies.⁵ If your content must pass a legal readability requirement, confirm which specific formula the regulation cites before you start writing.

When to use this

Compare the two scores above whenever you get a result that seems to contradict itself; a high Grade Level next to a moderate Reading Ease points to long sentences as the problem, while a low Reading Ease next to a moderate Grade Level points to dense vocabulary instead.

Examples

Your text scores: Reading Ease = 65, Grade Level = 8.5. What do these mean?

Reading Ease 65 = "Standard" (easily understood by 13–15 year olds). Grade 8.5 = approximately 8th–9th grade US student level. Both confirm the same audience: a general adult reader comfortable with moderate-length sentences and common vocabulary.

Text A scores 60 Reading Ease / Grade 8. Text B scores 60 Reading Ease / Grade 12. How is that possible?

Same Reading Ease means similar sentence length and word complexity. But Grade Level weightings differ slightly: longer sentences push Grade Level higher faster. Text B likely has very long but simple sentences, a pattern common in legal writing.

SMOG Grade and Other Readability Formulas

Your patient education brochure scores grade 8 on Flesch-Kincaid but grade 11 on SMOG. The scores disagree because the formulas measure different things. Knowing which formula your industry trusts prevents you from optimizing the wrong number and failing a compliance review.

SMOG Grade: designed for health literacy and patient education

SMOG (Simple Measure of Gobbledygook) was developed by G. Harry McLaughlin in 1969 specifically to evaluate the readability of health education materials. Unlike Flesch-Kincaid, which samples the full text, SMOG requires exactly 30 sentences: the first 10, the last 10, and 10 from the middle. It counts all words with three or more syllables in that 30-sentence sample and takes the square root. This construction makes SMOG specifically sensitive to polysyllabic medical terminology, which is why it became the standard formula for patient education materials and health literacy compliance.¹

Health communicators use SMOG to estimate how many years of education a reader needs to understand a document, and a common goal for general audiences is an 8th grade reading level or lower.² If your content must comply with a health literacy standard, confirm whether the specification references SMOG specifically, because other formulas produce different scores and substituting one for another may not satisfy the requirement.

SMOG thresholds in US healthcare guidelines

The American Medical Association and American Academy of Pediatrics both publish guidance recommending SMOG Grade 6 as the maximum for written patient instructions. CDC guidelines on health communication specify readability at a 6th-to-8th-grade level using a validated readability tool, with SMOG listed as a recommended option. For hospital patient education departments and pharmaceutical consumer information leaflets, these SMOG thresholds are operational requirements, not suggestions. Testing content against SMOG before publication is a compliance step in those contexts.

The thresholds exist because patient instructions fail their purpose when the reader cannot follow them. A discharge summary written above the patient's reading level is more likely to be misinterpreted, which is the specific harm these guidelines are designed to prevent. Treat the 6–8 grade band as a ceiling for any document a patient must act on without clinical support.

Coleman-Liau Index: character-based readability without syllable counting

Coleman-Liau calculates readability from characters per 100 words and sentences per 100 words, not from syllable counts. Its formula is: 0.0588 × (characters per 100 words) − 0.296 × (sentences per 100 words) − 15.8. This character-based approach eliminates the need for a syllable dictionary, making Coleman-Liau the most reproducible formula across different tools: two programs that count characters reach the same score, while two programs using different syllable dictionaries may diverge on Flesch-Kincaid. For automated content auditing systems processing thousands of documents, this consistency is a practical advantage.³

In practice, Coleman-Liau tends to score 1–2 grade levels higher than Flesch-Kincaid for the same text. Character count captures word length more precisely than syllable count: a three-syllable word with 12 characters scores higher than a three-syllable word with 8 characters. This sensitivity to character-per-syllable ratio makes Coleman-Liau more responsive to vocabulary complexity in content where words vary significantly in character length, such as compound technical terms common in engineering and legal writing.

Coleman-Liau in computational text analysis

Natural language processing and corpus linguistics researchers prefer Coleman-Liau over Flesch-Kincaid for large-scale analysis because it requires only character and sentence counts. Both are available from standard text processing libraries without additional lookup tables. If you are building a content auditing system that processes thousands of pages automatically, Coleman-Liau adds less computational overhead than any syllable-based formula. For manual readability checks on individual documents, the choice between formulas matters less; for batch processing pipelines, Coleman-Liau's calculation simplicity gives it a practical edge.

Automated Readability Index, Gunning Fog, and choosing the right formula

The Automated Readability Index (ARI) uses characters per word and words per sentence: 4.71 × (characters/words) + 0.5 × (words/sentences) − 21.43. Developed for the US Air Force in 1967, ARI tends to produce scores 1–2 grade levels higher than Flesch-Kincaid, particularly for short documents. It is most useful when you need a high-confidence score for a document too short for SMOG, which requires a minimum of 30 sentences. Gunning Fog counts only true complex words, excluding proper nouns and common inflected forms, which makes it less sensitive to named entities and more accurate for business writing dense with brand and company names.⁴

Matching each formula to its industry context

No single formula is right for every context. Flesch-Kincaid remains the most broadly recognized default. Use SMOG when content must meet a health literacy standard. Prefer Coleman-Liau for automated analysis pipelines where reproducibility matters most. For business writing training programs, Gunning Fog at 10–12 is the benchmark most corporate coaches reference, since the index treats 7 or 8 as the ideal and anything above 12 as very difficult for most readers⁵. After revising content, re-score it with multiple formulas so you keep patient copy inside SMOG limits rather than improving on one formula alone.

When to use this

Use this guide when Flesch-Kincaid alone does not give you enough insight, or when your audience cares about a specific formula (health literacy standards, for example, often require SMOG).

Examples

A paragraph from a medical website scores: Flesch-Kincaid Grade 14, but you do not know what that means relative to other formulas.

The same paragraph scores: Flesch-Kincaid Grade 14 | SMOG Grade 12 | Coleman-Liau Grade 15 | ARI Grade 13 | Gunning Fog 16. Consistently high scores across formulas confirm the text genuinely requires advanced reading ability and is not a quirk of one formula's calculation method. Use these results as a baseline and rewrite to bring all scores into a target range.

Patient education materials must comply with a SMOG Grade ≤ 8 requirement.

Rewrite until SMOG drops to 8 or below. Focus on replacing multi-syllable medical terms with plain equivalents, breaking long sentences, and adding analogies. Re-score after each revision to track progress.