Glossary · 3 · Understanding
Readability formula
Also known as: Readability index, Readability score
German: Lesbarkeitsformel
A readability formula is a mathematical formula that estimates the reading difficulty of a text from countable surface features, typically average sentence length and average word length in syllables or letters, such as the Flesch Reading Ease score (1948), the Flesch–Kincaid Grade Level (1975) or, for German, the Amstad adaptation of the Flesch score.
- Applied cognitive psychology
- Understanding
- Evaluation methods
In one sentence
A readability formula estimates reading difficulty from surface features such as sentence and word length, for example Flesch Reading Ease (1948).
Example
An editor's tool reports a Flesch–Kincaid grade of 14 for a troubleshooting topic. Instead of chopping sentences to reach a lower score, the writer tests the topic with two technicians and finds the real problem: an undefined fault code.
Rests on
- Working memory: long sentences and long words correlate with processing load, which is why the formulas use them as proxies.
- Long-term memory: frequent, short words are recognized faster; word length is a rough proxy for familiarity.
How it applies
- Screening, not judging: A formula can flag passages that may be hard and track a large body of text over time, for example in automated style checking.
- History in technical documentation: The Flesch–Kincaid Grade Level and the Automated Readability Index were recalculated in 1975 for US Navy enlisted personnel and technical training material (Kincaid, Fishburne, Rogers and Chissom).
- Other languages: Formulas are language-specific; the Flesch score was adapted to German by Toni Amstad (1978).
Evidence status
- Readability formulas measure surface features only. They ignore content, the reader's prior knowledge, coherence, layout, figures and retrieval aids. Redish (2000) summarized their technical weaknesses for technical documentation: developed for children's school books, not adult technical texts, blind to differences between readers and to the effects of content and layout.
- Rewriting a text to a target score can reduce clarity: splitting sentences removes the connectives that show relations, and replacing a precise term with a shorter one can make the text wrong.
- They do not apply to steps, tables, code, lists and fragments, which make up much of technical documentation.
Limits
- A score is never evidence that readers understand a text; test with readers instead (see Comprehensibility and Usability testing).
- No score shows that safety information is adequate.