In short
- To read fluently and in volume, a learner needs to know about 98% of the words on the page, which for most school learners means graded text.
- Novels and newspapers written for native speakers demand around 8,000 to 9,000 word families for that level of coverage.
- Authentic texts have a real place in short, teacher-supported lessons and with higher-level learners.
- In the studies reviewed here, simplified texts were read faster and understood better, while results for language learning itself are mixed.
Teachers are often told to use 'real' English: news articles, novels, web pages. They are also told to give pupils easy books and let them read. Both pieces of advice are sound, but they serve different purposes. This article sets out what the research on vocabulary coverage says, what authentic novels and newspapers demand of a reader, and what happened when researchers compared simplified and authentic texts directly. It ends with guidance by purpose and by level. 'Authentic' here means written for native speakers, not for learners.
Which should learners read?
For reading in volume, learners should read text in which they know nearly all the words. For most school learners, that means graded readers. Authentic texts, written for native speakers, suit short lessons with teacher support and learners at higher levels.
The choice depends on purpose. If the aim is fluent reading, many pages and a reading habit, choose graded text. If the aim is to study language closely, or to practise coping with real-world text, a short authentic text can do the job.
How many words must a reader know to follow a text?
Researchers call this vocabulary coverage: the share of the running words in a text that the reader already knows. Hu and Nation (2000) tested it directly. They gave 66 adult learners a 673-word story in four versions. In three of them, 5%, 10% or 20% of the words were replaced by nonsense words.
With 80% of the words known, no reader reached adequate comprehension. With 90% or 95% known, some did and most did not. The authors concluded that around 98% coverage is needed to read fiction without help. Their paper shows what each figure means on the page:
- 98% known: one unknown word in 50, about one in every five lines.
- 95% known: one unknown word in 20, about one in every two lines.
- 90% known: one unknown word in 10, about one in every line.
- 80% known: one unknown word in five, about two in every line.
Schmitt, Jiang and Grabe (2011) tested 661 learners in eight countries on two texts. They found no sudden threshold. Comprehension rose steadily as the share of known words rose. They judged 98% a more reasonable target than 95% for readers of academic texts.
Two cautions apply. Hu and Nation's readers were adults preparing for university, and they read one short text. And coverage is not everything: the authors note that grammar, reading experience and background knowledge also helped their readers. Even so, neither study suggests pupils read well when one word in ten is unknown.
How many word families do authentic texts demand?
Nation (2006) counted the vocabulary in novels and newspapers written for native speakers. He asked how many word families a reader needs in order to know 98% of the running words. A word family is a base word plus its close relatives, such as 'help', 'helps' and 'helpful'.
| Text type | Word families needed | Notes |
|---|---|---|
| Novels for native speakers | 8,000 to 9,000 | Five novels were analysed. The 2,000 most frequent word families covered about 88% of the words. |
| Newspapers | 8,000 to 9,000 | The 2,000 most frequent word families covered about 83%. Names made up about 4.5% to 6% of the words. |
| One graded reader written within 1,000 headwords | 682 in the whole book | A simplified classic of 10,578 words. Nation found few words outside the most frequent ones. |
Nation's figures show what this means for a learner who knows the 2,000 most frequent word families. In a novel, that learner knows about 88% of the words, plus the names. Roughly one word in ten is unknown, which is about one in every line.
Even a reader with 9,000 word families still meets one unfamiliar word in fifty, Nation adds. For estimates of vocabulary size by level, see How Many Words Do You Need at Each CEFR Level?
The case against simplified texts
The argument for authentic texts is serious. Gilmore (2007) reviewed the research on authenticity in language teaching. He reports a long-recognised gap between the language of textbooks and the language people really use. In one example, a textbook gave learners a distorted picture of how speakers make invitations.
Those who favour authentic texts usually argue three things:
- Simplifying a text can strip out the natural patterns, idiom and detail that learners need to meet.
- Learners who only read simplified text are not prepared for the texts they will meet outside school.
- Real texts can motivate, because they are about the real world.
Gilmore (2007) also describes a practical alternative: keep the text and vary the task. A class can read a real leaflet only to find three facts. Many researchers, he reports, believe learners at all levels can cope with authentic material if texts and tasks are carefully selected.
This case is strongest for lessons, where a teacher is present and the text is short. Reading in volume is a different case. Nobody is there to help, and the pupil must carry on alone for many pages.
What studies comparing simplified and authentic texts found
Crossley, Yang and McNamara (2014) tested 48 learners in Mexico, aged 15 to 24. The materials were authentic texts and versions of them simplified to beginning and intermediate levels. Each learner read nine texts, three of each kind. Beginning-level versions were read faster and understood better than intermediate or authentic versions.
The comprehension advantage held after the researchers controlled for language proficiency, reading proficiency and background knowledge. The speed advantage did not: once reading proficiency was counted, text type no longer made a significant difference to reading time.
The same authors describe how the texts differ. Simplified texts use more familiar words, are more cohesive and have simpler syntax. Higher cohesion means ideas are linked more clearly from one sentence to the next. In earlier analyses they summarise, authentic texts at intermediate level had more complex syntax and more logical connectors.
Beglar, Hunt and Kite (2012) followed 97 first-year Japanese university students for a year. The pleasure reading groups gained more reading speed than an intensive reading group, and comprehension stayed high. Reading simplified texts led to greater reading rate gains than reading unsimplified texts.
Gilmore (2007) summarises an older study of Japanese college students by Yano, Long and Ross. Simplified versions and elaborated versions both improved comprehension compared with the unmodified text. Elaboration means adding clarification and rephrasing to a text, not cutting it.
The evidence has gaps. Gilmore (2007) describes the results for language acquisition, as opposed to comprehension, as mixed. The participants in these studies were older teenagers and adults. None of the studies cited here followed pupils aged 8 to 14.
A good graded reader is literature written for learners
The word 'simplified' suggests a lesser copy of something real. Day and Bamford (1998) reject that picture. They call graded readers language learner literature, a genre like children's literature or young adult literature. Nobody calls a children's novel inauthentic because it was written for children.
On this view, authenticity lies in the reading as well as in the text. A pupil who laughs at a story, or reads on to find out who took the money, is reading for real. A pupil decoding a newspaper article word by word is not.
Quality varies, as with any books. A well-written graded reader has a plot worth following and language that sounds natural within its limits. For how the books are made and labelled, see What Are Graded Readers?
Which text for which purpose?
| Purpose | Best text type | Why |
|---|---|---|
| Reading in volume, in class or at home | Graded readers at or just below the pupil's level | Readers need to know about 98% of the words to read without help (Hu and Nation, 2000). |
| Building reading speed | Easy graded text | Simplified texts gave greater reading rate gains than unsimplified texts (Beglar, Hunt and Kite, 2012). |
| Close study of language in a lesson | A short text, graded or authentic, with teacher support | The teacher can explain unknown words, so lower coverage is workable. |
| Practising real-world reading tasks | A short authentic text with an easy task | The task can be simple even when the text is not (Gilmore, 2007). |
| Preparing for exams that use unsimplified passages | Authentic or exam-style texts, in short supported sessions | Pupils need to meet the kind of text they will be tested on. |
| Free reading for advanced learners | Authentic books chosen by the pupil | Novels need 8,000 to 9,000 word families for 98% coverage (Nation, 2006). |
Classroom example (invented for illustration)
A B1 class is reading a graded detective story at home, one chapter a night. Nobody uses a dictionary.
On Thursday the teacher brings a real 150-word news report about a museum theft. Pupils underline the words they do not know. Most underline ten or more.
The task is easy: find who, where and when. Then the class studies four of the words together.
The graded story builds fluency and habit. The news report is fifteen minutes of supported work with real text. Neither replaces the other.
Guidance by level
- A1 and A2. Use graded text for all independent reading. Use authentic material only as very short items with a simple task: a sign, a menu, a label.
- B1. Graded readers remain the main diet. Add short authentic texts in class, with support and an easy task.
- B2. Pupils can begin authentic books they choose themselves, alongside upper-level graded readers. Check a page first and expect slower reading.
- C1 and above. Authentic text for most reading.
These bands are a teaching judgement built on the coverage figures, not a tested rule. No study cited here assigns text types to CEFR levels. Check each pupil with a page of the actual book, as How to Match Students to the Right Graded Readers explains.
The difference between reading in volume and close study is set out in Extensive vs. Intensive Reading: What's the Difference? The guide to extensive reading shows how both fit into a school week.
How iRead handles this
In iRead, every story is a graded reader levelled to the CEFR, from A1 to C2. A2 stories are short and simple, on familiar topics with high-frequency vocabulary. B1 brings longer narratives read independently, and C1 brings sophisticated texts with idiomatic language. Many stories have an audio read-along synced to the text. Schools can also build school packs from their own books.
The core principle: Match the text to the purpose. For pages and fluency, choose text the pupil can almost fully understand. For close study, use a short harder text with your support.
See it in iRead: Every iRead story is a graded reader levelled to the CEFR, from A1 to C2. Schools can use packs from the iRead library or build school packs from their own books.
See how iRead worksKey takeaway
Give pupils graded text for the reading they do alone and in volume. Bring in authentic text in short, supported lessons, and hand over to it fully as pupils reach higher levels.
Frequently asked questions
Should English learners read authentic texts?
Yes, but for the right purpose. Short authentic texts work in lessons where the teacher supports the reading and sets an easy task. For reading alone and in volume, most school learners need graded text, because authentic books contain too many unknown words.
What is the difference between a graded reader and an authentic text?
An authentic text was written for native speakers: a novel, a news article, a web page. A graded reader was written or adapted for language learners, with vocabulary and grammar kept within a set level so that the learner can read it without help.
What percentage of words must a learner know to understand a text?
Hu and Nation (2000) concluded that readers need to know about 98% of the running words to follow fiction without help. Schmitt, Jiang and Grabe (2011) found no sharp threshold: comprehension rose steadily as the share of known words rose.
How many words do you need to read a novel in English?
Nation (2006) estimated that 8,000 to 9,000 word families are needed to know 98% of the words in novels written for native speakers. Newspapers need a similar number. Even then, about one word in fifty will be unfamiliar.
Do simplified texts harm language learning?
The studies reviewed here do not show harm. Simplified texts were understood better and gave greater reading rate gains. Gilmore (2007) reports mixed results for language acquisition, so the honest answer is that long-term effects on learning are less certain.
When can learners move from graded readers to authentic books?
When a pupil can read a page of the authentic book and meet only one or two unknown words. Try from around B2, but it varies by pupil and by book. Check a page of the actual book instead of relying on the level.
Sources
- Hu, M. H.-C., & Nation, P. (2000). Unknown vocabulary density and reading comprehension. Reading in a Foreign Language, 13(1), 403-430. doi.org/10.64152/10125/66973
- Schmitt, N., Jiang, X., & Grabe, W. (2011). The percentage of words known in a text and reading comprehension. The Modern Language Journal, 95(1), 26-43. doi.org/10.1111/j.1540-4781.2011.01146.x
- Nation, I. S. P. (2006). How large a vocabulary is needed for reading and listening?. The Canadian Modern Language Review, 63(1), 59-82. doi.org/10.3138/cmlr.63.1.59
- Gilmore, A. (2007). Authentic materials and authenticity in foreign language learning. Language Teaching, 40(2), 97-118. doi.org/10.1017/S0261444807004144
- Crossley, S. A., Yang, H. S., & McNamara, D. S. (2014). What's so simple about simplified texts? A computational and psycholinguistic investigation of text comprehension and text processing. Reading in a Foreign Language, 26(1), 92-113. doi.org/10.64152/10125/66686
- Beglar, D., Hunt, A., & Kite, Y. (2012). The effect of pleasure reading on Japanese university EFL learners' reading rates. Language Learning, 62(3), 665-703. doi.org/10.1111/j.1467-9922.2011.00651.x
- Day, R. R., & Bamford, J. (1998). Extensive Reading in the Second Language Classroom. Cambridge University Press. www.cambridge.org/9780521568296