Readability Checker

“The cat sat on the mat.” 6 words, 1 sentence, 6 syllables. Reading ease = 206.835 − 1.015 × 6 − 84.6 × 1 = 116.1, and Flesch–Kincaid = 0.39 × 6 + 11.8 × 1 − 15.59 = −1.45. Very simple text can score above 100 and below grade 0; the scales have no hard ends.

The Gettysburg Address. 271 words (“battle-field” counts as one), 10 sentences and 367 syllables: 27.1 words per sentence and 1.354 syllables per word. Reading ease = 206.835 − 1.015 × 27.1 − 84.6 × 1.354 = 64.8, an 8th–9th grade level on Flesch’s table. The five grade formulas give Coleman–Liau 8.0, Flesch–Kincaid 11.0, SMOG 11.2, ARI 12.1 and Fog 13.3, so the consensus (the middle one) is 11.2. The short words keep the reading ease high; the long sentences push the grades up.

The common mistake: comparing scores from tools that split sentences differently. Flesch’s 1979 rules count a dash, colon or semicolon as the end of a sentence. Lincoln’s last sentence is full of dashes, so the same speech becomes 17 sentences, Flesch–Kincaid drops from 11.0 to 6.6 and reading ease rises from 64.8 to 76.1. Neither number is wrong; compare texts only under the same rule (switch it under “Sentences end at”).

Flesch Reading Ease (1948). Rudolf Flesch developed his formula in the early 1940s for his Columbia University doctorate and published it in the Journal of Applied Psychology in 1948 (“A new readability yardstick”, vol. 32, pp. 221–233). In How to Write Plain English he gives it as: multiply the average sentence length by 1.015 and the average word length in syllables by 84.6, add them, and subtract the sum from 206.835. His table puts 90–100 at 5th grade, 80–90 at 6th, 70–80 at 7th, 60–70 at 8th and 9th, 50–60 at high school, 30–50 at college and 0–30 at college graduate. He measured comics at 92, Reader’s Digest at 65, Time at 52 and the Harvard Law Review at 32.

Gunning Fog (1952). Robert Gunning’s The Technique of Clear Writing adds the average sentence length to the percentage of words of three or more syllables and multiplies by 0.4. He told readers not to count proper names, combinations of short easy words (“bookkeeper”) or verb forms made three syllables by -ed or -es (“created”), and called 13 or more the danger line. This checker applies those exclusions to capitalised words, hyphenated compounds and -ed/-es forms; untick the box to count names too.

Automated Readability Index (1967). R. J. Senter and E. A. Smith built ARI at the Aerospace Medical Research Laboratories around an IBM Selectric typewriter fitted with counters for strokes, words and sentences, so no syllables were needed. Fitted to graded readers for grades 1 to 7, their equation is grade = 0.50 × words per sentence + 4.71 × strokes per word − 21.43 (multiple R 0.98). Their “strokes” were every printing key, punctuation included; like most software, this checker counts letters and digits only.

SMOG (1969). G. Harry McLaughlin fitted his formula to the 390 passages of the McCall-Crabbs Standard Test Lessons in Reading, taking as the target the grade of readers who answered every question correctly. The quick method: count the words of three or more syllables in 30 sentences, take the square root, add 3. The regression form used here, 1.0430 × √(count) + 3.1291, scales the count to 30 sentences. McLaughlin put the standard error at about 1.5 grades.

Coleman–Liau and Flesch–Kincaid (1975). Meri Coleman and T. L. Liau designed their index for optical scanners: letters per 100 words and sentences per 100 words, no syllables. The same year J. Peter Kincaid and colleagues recalculated Flesch, Fog and ARI for the US Navy from the reading tests of 531 enlisted personnel; their version of Flesch, 0.39 × words per sentence + 11.8 × syllables per word − 15.59, is the Flesch–Kincaid grade level.

Why tools disagree. The formulas are fixed, but the counts are not. Syllables are the weak point: on 96,000 words of modern encyclopedia text this checker’s counter (rules plus a list of exceptions) agrees with the CMU Pronouncing Dictionary on 99.5% of words, while a plain vowel-group counter agrees on 91%. On words of three or more syllables, which drive Fog and SMOG, the agreement is 99.8% against 96.6%. Sentence rules matter even more: dashes, colons, bullet lists and abbreviations such as “Mr.” can change the sentence count by half. That is why the consensus here is the middle of five scores, and why every count is shown in Show Work.

This checker scores any English text on six published readability formulas (Flesch Reading Ease, Flesch–Kincaid, Gunning Fog, SMOG, Coleman–Liau and the Automated Readability Index) with the constants from the original papers, and shows the middle of the five grade levels as a consensus. It marks sentences over your length limit and words of three or more syllables, so you can see where the difficulty comes from, and shows every count and formula in Show Work.

The text is analysed in your browser as you type and is never sent anywhere. It copes with a whole book: the 130,000 words of Pride and Prejudice score in under a tenth of a second on a laptop, and the marked copy shows the first 20,000 words. It is built for teachers matching texts to a class, students checking essays, and anyone writing letters, leaflets or instructions for the public.

Related tools: Grade Calculator, EZ Grader, and Rubric Score to Grade Converter.