Readability score checker

The six formulas

All six turn two simple measurements into a score: how long the sentences are, and how long the words are. They differ in how they measure word length and in the texts they were calibrated against. In the formulas below, W is words, S sentences, Sy syllables.

Flesch Reading Ease (Rudolf Flesch, 1948)

206.835 − 1.015 × (W/S) − 84.6 × (Sy/W). A score from about 0 to 100 where higher is easier; very simple text can go above 100 and very dense text below 0. It is the only one of the six that is not a school grade.

Flesch-Kincaid Grade Level (Kincaid and colleagues, 1975)

0.39 × (W/S) + 11.8 × (Sy/W) − 15.59. The same two inputs as Reading Ease, re-weighted to give a US grade. Developed for the US Navy to check training material, and now the level most word processors report.

Gunning Fog Index (Robert Gunning, 1952)

0.4 × [(W/S) + 100 × (complex words / W)], where a complex word has three or more syllables. Following Gunning's rules, a word does not count as complex if it only reaches three syllables because of an -es, -ed or -ing ending. His other exclusions, proper nouns and familiar compounds, are not applied here, so Fog can read slightly high on text full of names.

SMOG (G. Harry McLaughlin, 1969)

1.043 × √(polysyllables × 30/S) + 3.1291. Counts words of three or more syllables, scaled to a 30-sentence sample. SMOG was calibrated against complete comprehension rather than partial, so it usually reads a grade or two higher than the others. It is widely used for health information for that reason.

Coleman-Liau Index (Meri Coleman and T. L. Liau, 1975)

0.0588 × L − 0.296 × S₁₀₀ − 15.8, where L is letters per 100 words and S₁₀₀ is sentences per 100 words. It counts letters instead of syllables, which a computer can do exactly.

Automated Readability Index (Senter and Smith, 1967)

4.71 × (letters/W) + 0.5 × (W/S) − 21.43. Also letter-based, and designed for the US Air Force so that the counting could be done by machine as text was typed.

A worked example

The sample passage in the box above has 105 words in 7 sentences, with 144 syllables. That is 15.0 words per sentence and 1.37 syllables per word. Flesch Reading Ease is then 206.835 − 1.015 × 15.0 − 84.6 × 1.37 = 75.6, which is in Flesch's "fairly easy" band.

The grade-level formulas put the same passage anywhere from grade 6.3 to grade 10.9. SMOG gives the highest figure, partly because it expects complete comprehension and partly because a seven-sentence sample has to be scaled up to its 30-sentence design. That spread is the real answer: the passage suits a reader at roughly middle-school level or above, and no formula can be more precise than that.

Reading Ease score bands

Flesch described his scale in bands. The school levels attached to them are the conventional ones, not a guarantee about any particular reader.

What readability scores cannot tell you

Every formula here counts surface features. None of them reads. That leads to predictable blind spots:

• Meaning. Shuffle the words in each sentence and every score stays exactly the same. Short words can express hard ideas, and a well-built long sentence can be easier than three choppy ones.
• Familiarity. "Television" has four syllables and every reader knows it; "gist" has one and many do not. Formulas treat length as a stand-in for difficulty, which works on average and fails on individual words.
• Layout. Headings, lists and short paragraphs make a document much easier to use, and none of that changes a score. Bullet points without full stops may even count as one long sentence if they are on one line.
• Syllable counting. Syllables here are estimated from spelling, which is right for most words and off by one for some. On a passage of a few hundred words the errors mostly cancel out, but on a single sentence they can move the score noticeably.

Use the scores to spot text that is drifting too dense, and to compare two drafts of the same document. Writing to hit a number, by chopping every sentence in half, tends to make text worse.