Gunning Fog Index
A readability index that maps a text to a US-grade scale using sentence length and the percentage of complex words, defined as words of three or more syllables. Developed by Robert Gunning, a corporate writing consultant, and published in his 1952 book The Technique of Clear Writing. The name refers to the fog of unnecessarily complicated prose Gunning's clients produced in business writing; the index is a tool for identifying it.
The formula
The output is a US grade level. A Fog of 12 means a passage broadly readable by an average twelfth-grader. Below 8 is very easy; 8–12 is plain English; above 17 is the graduate-level zone where Gunning argued business writing should never be.
The complex-word rules, and why most tools ignore them
Gunning's original specification for complex words is more restrictive than the three-syllable threshold suggests. The 1952 rules exclude:
- Proper nouns (
Baltimore,Mrs. Madison,Microsoft). - Compound words built from short, familiar pieces (
bookkeeper,sunflower,butterfly). - Words made polysyllabic only by common inflectional suffixes (
-ed,-es,-ing). Socreatedandtrespassesdo not count even though both have three syllables.
Almost no software implementation applies the full exclusion list. Most count every 3+-syllable token as complex. The resulting scores read systematically higher than Gunning's original specification would produce, often by 1–2 grade levels on technical or academic prose. This is one of the most reliable inter-tool drift sources in Text Metric Implementation Variance.
Gunning's exclusions encode a real psycholinguistic insight: the cognitive load of a polysyllabic word depends on whether it is intrinsically polysyllabic (circumlocution, phenomenology) or polysyllabic only because of regular morphology (walking, rebuilding). Strict implementations are closer to the construct; loose implementations are easier to write.
Strengths and limits
Fog's strength is interpretability for writers. The two predictors map cleanly to the two pieces of advice business-writing manuals have given for a century — shorten your sentences, simplify your words — and a Fog reduction is directly traceable to changes a writer can make. Below the surface, Fog correlates strongly with FKGL (typically r > 0.9) on natural prose; the two indices rank texts almost identically.
The limits are the standard surface-feature limits. Fog does not see vocabulary frequency, topical demand, cohesion, or rhetorical organisation. A medical-research abstract can score Fog 14 because it uses Latinate vocabulary in short sentences, while a folk tale can score Fog 11 because it uses simple words in long sentences. For CEFR alignment Fog is one of three or four formulas to combine, never to use alone.
When to prefer Fog over FKGL or FRE
The case for Fog is professional-writing pedagogy. Its predictors map directly onto the two largest moves a business or academic writer can make to improve clarity, and the grade-level output is communicative without being bound to the FKGL formula's idiosyncratic coefficients. For ELT writing-skills work — particularly business and academic writing courses — Fog is the readability number most often built into rubrics because students can act on it directly.
For test passage sourcing, SMOG is generally preferred over Fog at the very-difficult end of the scale (its square-root scaling handles dense text more stably), and New Dale-Chall is preferred when vocabulary-frequency sensitivity matters more than sentence-length sensitivity.
Key References
- Gunning, R. (1952). The Technique of Clear Writing. McGraw-Hill.
- Gunning, R. (1968). The Technique of Clear Writing (revised edition). McGraw-Hill.
- DuBay, W. H. (2004). The Principles of Readability. Impact Information.
See Also
- Readability: the umbrella construct
- Flesch Reading Ease / Flesch-Kincaid Grade Level: surface-feature formulas Fog correlates with at r > 0.9
- SMOG: the simpler-to-compute square-root alternative for very-difficult text
- Dale-Chall Readability Formula: the frequency-list alternative when word difficulty matters more than syllable count
- Text Metric Implementation Variance: the complex-word exclusion rules and why most implementations skip them