Flesch-Kincaid

The readability formula behind the Headline Analyzer's grade-level score — where it came from, how it's calculated, and where it breaks down.

What it is

Flesch-Kincaid is a family of readability formulas that estimate how difficult a piece of English text is to read, expressed either as a 0–100 ease score or as a U.S. school grade level. The two most common variants are the Flesch Reading Ease score (higher means easier to read) and the Flesch-Kincaid Grade Level (an estimate of the U.S. school grade a reader would need to have completed to understand the text on a first read).

Where it came from

The underlying Flesch Reading Ease formula was developed by Rudolf Flesch in 1948. In 1975, J. Peter Kincaid and colleagues adapted it into the grade-level version specifically for the U.S. Navy, to gauge whether technical training manuals matched the reading level of enlisted personnel. The original research report, Derivation of New Readability Formulas for Navy Enlisted Personnel, is freely available through the Internet Archive.

How it's calculated

Both formulas are built from the same two raw signals: average sentence length (words per sentence) and average word length (syllables per word). The grade-level formula is:

Grade Level = 0.39 × (words ÷ sentences) + 11.8 × (syllables ÷ words) − 15.59

Longer sentences and longer (more syllable-heavy) words both push the estimated grade level up. There's no dictionary lookup or semantic understanding involved — it's purely structural, which is both the formula's strength (fast, consistent, language-agnostic to compute) and its weakness (see below).

Where it breaks down

The formula was designed and validated against multi-sentence prose, not single short phrases. Applied to something as short as a headline, the sentence-count denominator can produce unstable or even negative results — a five-word phrase with one long word can swing the estimate sharply. It also can't distinguish a genuinely complex word from a long-but-familiar one; "elephant" and "epistemological" score identically for syllable count despite being nothing alike in actual difficulty.

How the Headline Analyzer uses it

The Headline Analyzer's readability sub-score is Flesch-Kincaid-based, but adjusted for the fact that headlines aren't prose: unlike long-form writing, a headline is never penalized for being simple. There's no such thing as a headline that reads "too easily." The score only costs points when the estimated grade level climbs unusually high — the one direction where the formula's signal is still meaningful even on short text.

Source

Kincaid, J.P., Fishburne, R.P., Rogers, R.L., & Chissom, B.S. (1975). Derivation of New Readability Formulas for Navy Enlisted Personnel. Research Branch Report 8-75, U.S. Naval Air Station, Memphis — freely available via the Internet Archive.

Back to the full glossary — 200 terms covering case conversion, style guides, and text tools.