A B1 reader opens The Great Gatsby on the Taipei MRT and meets 11 unknown words before the train reaches the next station. Four minutes go to a dictionary, the story disappears, and page two waits until Sunday. For fluency practice, an English book should feel easy enough to keep the reader inside the story for several pages at a time. This argument is for a university student who can pass an intermediate reading test yet still translates line by line. A short graded reader may look modest beside an American novel. It can produce the volume and continuity that reading fluency requires.
Difficulty changes the activity
Close reading and extensive reading have different jobs. Close reading can justify a dense paragraph, dictionary work, grammar analysis, and ten minutes spent on one sentence. Extensive reading asks the learner to follow meaning across many pages, meet familiar language again, and make word recognition less effortful.
One text cannot maximise both jobs at once.
When unknown words arrive every line, attention keeps leaving the message. The learner searches, translates, copies a definition, and reconstructs the sentence. That can be useful vocabulary study. It supplies poor conditions for practising fast recognition of words the learner already knows.
Marcella Hu and Paul Nation manipulated vocabulary coverage in a 633-word narrative read by 66 learners. Versions contained 80, 90, 95, or 100 percent known words, with nonsense words replacing unknown items. No learner at 80 percent reached the study's adequate-comprehension criterion. Scores varied widely at 90 and 95 percent. The authors used a regression model to estimate that around 98 percent coverage might be needed for most learners to understand adequately without assistance. That figure comes from a regression estimate. The experiment used one short narrative and invented unknown forms. A later replication by Benjamin Kremmel and colleagues added a 98 percent condition, a second genre, revised questions, and a larger Sri Lankan sample. Their results showed that the relation between unknown-word density and comprehension changed with genre and response format.
Background knowledge and reading skill also matter.
The 98 percent figure offers a useful warning about text load. It supplies no law for every reader and book.
Easy pages accumulate
David Beglar, Alan Hunt, and Yuriko Kite followed 97 first-year Japanese university students for one academic year. Three pleasure-reading groups were compared with an intensive-reading group. All three pleasure-reading groups increased reading rate more, and the two groups that read the most gained more than the pleasure-reading group that read the least, while comprehension stayed high from pretest to posttest. Learners using simplified texts also made larger rate gains than those reading unsimplified texts. Selection and motivation may explain part of the difference between groups. The useful scale comes from books and months: one book every two weeks or more was reported as the strongest pace for most learners in that sample.
Vocabulary growth follows prior knowledge
Stuart Webb and Anna Chang tracked 60 learners reading 20 graded readers across 37 weeks. The researchers grouped learners by their starting vocabulary knowledge and tested 100 target words from each of two sets of books. Immediate relative gains were 63.18 percent for the higher-knowledge group, 44.64 percent for the middle group, and 28.12 percent for the lower group. Three months later, much of the knowledge remained, though delayed relative gains ranged widely from 21.05 to 59.01 percent. The result makes a blunt point about incidental learning: learners get more from surrounding text when they already know enough of it. Reading alone may be too slow a route for a beginner who needs the first 1,000 frequent word families.
Direct vocabulary study, teacher explanation, and listening can support the books.
In another study, Webb and Chang compared 61 Taiwanese learners in extensive reading with audio support and a course-book approach. The extensive group read and listened to ten graded readers. Relative gains on 100 target words reached 44.06 percent immediately and 36.66 percent three months later. Frequency and distribution of individual target words did not show clear relationships with learning in that dataset, which cautions against promising that seeing a word a fixed number of times guarantees acquisition. Reading volume and vocabulary study can share the week. Choc's article on learning outside the classroom explains why class hours alone leave too little contact with English. Easy books give that outside time a sustainable unit: a chapter on Tuesday, another on Thursday, and a complete story before the following Monday.
Choose by observed reading behaviour
A usable fluency book passes four practical checks.
- The learner can read three pages without reaching for a dictionary.
- After 17 minutes, they can explain the main event and one unresolved detail.
- Dialogue sounds clear enough to read aloud without decoding every word first.
- The next chapter feels more attractive than the phone notification waiting beside it.
These checks diagnose the individual reader. A publisher label cannot do that alone. A Level 4 reader can be easy for one student and dense for another. Topic knowledge can make a book about semiconductor fabrication manageable even when its vocabulary profile looks difficult. A familiar plot can also carry a learner through older syntax. I would let the learner abandon a book after eight pages. Finishing every choice turns selection into punishment and hides useful evidence about level, genre, or interest. The replacement should become easier or more compelling for a named reason.
There is a place for hard books. Literature students need contact with original style. Exam candidates need timed work on texts they cannot fully control. A reader who loves baseball may tolerate a demanding Shohei Ohtani biography because prior knowledge fills gaps. These are separate reading purposes. For tomorrow's MRT ride, leave The Great Gatsby at home. Take the 64-page graded mystery, put the dictionary away for one chapter, and see whether Zhongxiao Fuxing arrives before the culprit does.