How to Measure the Reading Level of Your Writing (and What the Scores Mean)

Reading level is one of those concepts that is widely cited and widely misunderstood. The "Flesch-Kincaid Grade 8" number on a piece of writing does not mean "8th graders will read this" — it means something more specific and more useful, but also more limited, than most people assume. This post covers the six most common formulas, the differences, the math, the right one for the right audience, and how to run all six in your browser without uploading your draft to anyone.

What the formulas actually measure

Every readability formula reduces to two numbers: the average sentence length and the average word complexity (usually measured by the average number of syllables per word). The formulas combine these two numbers in different ways to produce a single "reading level" score. The combination is what makes the formulas different, and what makes the choice of formula matter for different audiences.

The implicit assumption: longer sentences and longer words are harder to read. This is broadly true — most readers process text in chunks, and longer chunks take more working memory to hold. But the assumption ignores many other factors: the reader's background knowledge, the topic, the formatting, the visual layout, the supporting images. A well-formatted article with short paragraphs and clear headings reads easier than its raw formula score suggests, and a dense academic paper reads harder than its score suggests.

For the practical purpose of "is this too dense for my audience?", the formulas are useful. For the academic purpose of "what is the actual cognitive load of this text", they are not. Use them for the first, not the second.

The six formulas

1. Flesch Reading Ease

One of the oldest (1948) and most widely used. Score 0-100, higher is easier. The formula:

FRE = 206.835 − 1.015 × (words/sentences) − 84.6 × (syllables/words)

Interpretation bands:

• 90-100: Very easy. 5th-grade reading level. Comic books, children's books.
• 80-90: Easy. 6th grade. Conversational writing, popular magazines.
• 70-80: Fairly easy. 7th grade. General web content, most newspapers.
• 60-70: Standard. 8th-9th grade. General adult content, Wikipedia average.
• 50-60: Fairly difficult. 10th-12th grade. Quality magazines, academic content.
• 30-50: Difficult. College. Scientific papers, professional journals.
• 0-30: Very difficult. College graduate. Legal contracts, dense academic text.

For most web content aimed at a general audience, 60-70 is the right band.

2. Flesch-Kincaid Grade Level

Same underlying signals (sentence length and syllables per word) as Flesch Reading Ease, but the output is a US grade level instead of a 0-100 score. The formula:

FKGL = 0.39 × (words/sentences) + 11.8 × (syllables/words) − 15.59

Score 8.0 means the text is estimated to be readable by an 8th-grader. For most web content, 7-9 is the right range. Above 12 starts to feel academic; below 6 starts to feel condescending for adult audiences.

3. Gunning Fog

Designed for business and journalism, gives a "years of education" score. The formula counts complex words (3+ syllables, excluding proper nouns, compound words, and common suffixes like -es, -ed, -ing):

Fog = 0.4 × (words/sentences + 100 × complex_words/words)

Score 12 means 12 years of education (i.e. senior in high school). For most general content, 8-12 is the right range. Above 14 starts to feel academic.

4. SMOG (Simple Measure of Gobbledygook)

Designed for healthcare content, gives a "years of education" score. The formula is similar to Gunning Fog but with a different weighting:

SMOG = 1.043 × sqrt(complex_words × 30 / sentences) + 3.1291

The notable feature is that SMOG requires at least 30 sentences to be accurate. For shorter texts, the score is unreliable. The Reading Level Calculator shows SMOG with a warning when the text is too short.

5. Coleman-Liau

Different from the other five: it does not use syllables at all. Instead, it uses the average number of characters per 100 words. The advantage is that it can be computed without a syllable-counter, which is faster and more reliable.

CLI = 0.0588 × L − 0.296 × S − 15.8

where L is the average number of characters per 100 words, and S is the average number of sentences per 100 words. The result is a US grade level. Useful as a cross-check against the syllable-based formulas; if they disagree, the text is unusual in some way (probably in its average word length).

6. Automated Readability Index (ARI)

Also uses characters, not syllables, and produces a US grade level:

ARI = 4.71 × (characters/words) + 0.5 × (words/sentences) − 21.43

Like Coleman-Liau, ARI is fast to compute and gives a result that usually tracks the syllable-based formulas within a grade or two.

Which one to use

For most uses, the right answer is to look at all six and treat them as a range. The Reading Level Calculator shows all six side by side, so you can see the consensus and the outliers.

• General web content (blog posts, marketing copy, product pages). Aim for Flesch Reading Ease 60-70 (Standard band) and Flesch-Kincaid Grade 7-9. The other formulas should be in a similar range.
• News and journalism. Aim for Flesch-Kincaid Grade 8-10. The New York Times averages around grade 10-11; the Wall Street Journal averages around grade 11-12. General-audit news sites are usually grade 9-10.
• Academic and scientific. Grade 13-16 is the norm. Trying to bring these below grade 12 usually means losing the precision that the content needs.
• Healthcare content for patients. Aim for grade 6-8. The American Medical Association and the National Institutes of Health both recommend grade 6-8 for patient-facing content. The Writing Style Analyzer is a useful companion here — it highlights the passive voice, long sentences, and complex words that push the grade up.
• Children's content. Grade 3-5 is the standard. The Flesch Reading Ease should be 80+.

What to do with the score

The score is a diagnostic, not a target. Use it to identify what is making the text dense:

• High score, but the sentences feel long. The sentence length is the problem, not the word complexity. Break long sentences into shorter ones.
• High score, but the words feel academic. The word complexity is the problem, not the sentence length. Replace multi-syllable words with shorter equivalents ("use" instead of "utilize", "show" instead of "demonstrate").
• High score, but the content is hard to follow for other reasons. The formula is not capturing the issue. Look at the structure (headings, paragraphs, lists), the formatting, the supporting images, and the order of the argument. The formula is a coarse signal; the actual readability is a multi-dimensional property.

The Writing Style Analyzer is a useful companion. It highlights the specific sentences and phrases that are dense, with rules for adverbs, passive voice, complex words, and long sentences. The combination of "here is the overall score" + "here are the specific sentences to fix" is the practical workflow.

Common limitations of the formulas

The formulas all share the same limitations, and they matter:

• Topic difficulty is not counted. A piece about a familiar topic (e.g. cooking) reads easier than its formula score suggests. A piece about an unfamiliar topic (e.g. blockchain) reads harder than its formula score suggests. The formula sees only the words and the sentences, not the prior knowledge the reader brings.
• Formatting is not counted. A well-formatted article with headings, lists, and short paragraphs reads easier than a wall of text with the same words. The formula sees only the words and the sentences.
• Voice and tone are not counted. A conversational piece ("Let's look at how this works") reads easier than a formal piece ("The mechanism is examined in the following analysis") with the same formula score. The formula does not distinguish.
• Syllable counting is imperfect. Most syllable counters are heuristic. "Every" is two syllables by one rule and three by another. "Business" is two syllables. The result is that the syllable-based formulas can be off by a few points for individual words, which compounds over a long text.

These limitations do not make the formulas useless — they make the formulas a useful first signal, not a final answer. The right workflow is to use the formula to identify potential issues, then read the text yourself to confirm.

How to use the Reading Level Calculator

1. Open the Reading Level Calculator.
2. Paste your text into the input field.
3. The six scores appear in real time as you type, with the interpretation for each one (band for Flesch Reading Ease, grade level for the others).
4. Look for the formula that matches your audience (most general web content aims for Flesch-Kincaid Grade 7-9).
5. For a deeper view, paste the text into the Writing Style Analyzer to see the specific sentences and phrases that are pulling the score up.

The whole analysis runs in the browser. The text is in your input field, the calculations are in JavaScript, and the result is rendered locally. Nothing is sent to a remote service, nothing is logged, and the draft does not need to leave your device.

A short pre-publish checklist for readability

1. Run the Reading Level Calculator. Check that the score matches the audience.
2. Run the Writing Style Analyzer. Fix the highlighted sentences — they are the ones pulling the score up.
3. Check the Sentence Length Checker for any sentences over 30 words. Break them up.
4. Read the text aloud. If you stumble, the reader will too. The formulas are a useful signal, but the human read-aloud test is the final check.

The right combination of formulas and human reading gives you text that is both accessible and accurate. The formulas catch the structural issues; the human read catches the voice and flow. Both are needed.