Listening texts demand fewer word families than reading texts to reach the same coverage figure. The pattern, established by Nation (2006) and reinforced by Webb and Rodgers' television and movie corpora, has direct consequences for materials design: a learner population that is far below the reading threshold for authentic written texts can often reach 95–98% coverage on colloquial spoken material with substantially less vocabulary.
Nation profiled spoken English (transcripts of unscripted conversation, Friends episodes, and similar colloquial registers) against fourteen 1,000-word-family lists from the British National Corpus and computed the vocabulary needed to reach 95% and 98% coverage. The headline contrast: written text needs 8,000–9,000 word families for 98% coverage; colloquial spoken English needs around 6,000–7,000. At 95% coverage, written text needs 4,000–5,000 word families; colloquial spoken English needs around 3,000.
The gap reflects the lexical character of informal speech: heavy reliance on the most frequent few thousand families, formulaic chunking, recycled topics, and a shallower long tail of rare items. Written prose, especially academic and literary writing, draws on a far broader low-frequency vocabulary.
Webb and Rodgers (2009a) profiled 88 television programmes totalling 264,384 running words across multiple genres. Knowledge of the most frequent 3,000 word families plus proper nouns and marginal words gave 95.45% coverage of the corpus; 7,000 word families gave 98.27%. Across genres the 95% target ranged from 2,000 to 4,000 families.
Webb and Rodgers (2009b) profiled movie scripts and reported a similar pattern, with 3,000 word families reaching 95% coverage and 6,000–7,000 reaching 98%. Television and movies sit close to the colloquial spoken benchmark from Nation 2006.
Coverage targets for spoken input may also be lower than for reading. Van Zeeland and Schmitt (2013) manipulated coverage of four informal narrative listening passages for 36 native and 40 non-native listeners. Most participants comprehended adequately at 90% coverage, with comprehension stabilising and variance dropping at 95%. The authors proposed 95% as a working target for informal listening, against the 98% target dominant in reading research, with the lower figure attributed to the redundancy and prosodic support of spoken delivery.
Academic spoken English shifts the figures upward. Dang and Webb (2014), profiling the British Academic Spoken English (BASE) corpus, reported that 2,000 word families reach 90% coverage of academic lectures and seminars, 4,000 reach 95%, and 8,000 reach 98%. Academic listening sits close to academic reading at the upper end and clearly above colloquial spoken English, reflecting the technical and sub-technical vocabulary load of university teaching.
Listening can be pitched lower than reading at the same coverage target. A learner with 3,000 word families is comfortably positioned for 95% coverage of mainstream television and informal conversation but well short of 95% coverage of authentic written prose. Extensive listening curricula and graded listening materials exploit this gap by routing learners through colloquial spoken input at coverage levels that authentic reading would not yet support.