Extensive Reading for Language Learning: What the Research Found

Multiple meta-analyses — the most recent covering 74 studies and thousands of learners — confirm that reading level-appropriate material regularly improves vocabulary, comprehension, reading speed, and overall language ability. But the picture is more nuanced than most language learning sites will tell you.

This article covers what the research actually found, where the evidence is strong, where it is weaker than often claimed, and what it means in practice for learners.

What is extensive reading?

Extensive reading (ER) means reading a large amount of material at or slightly below your current level — books, stories, articles — where you understand enough to follow the text without constantly looking up words. The goal is volume and enjoyment, not intensive analysis.

This contrasts with intensive reading, where you work through shorter, harder texts with a dictionary, focusing on grammar and vocabulary in detail.

The idea behind ER is simple: the more language you absorb in context, the more vocabulary, grammar patterns, and fluency you develop — without drills.

What did 74 studies find?

The most comprehensive meta-analysis to date is Sangers et al. (2025), published in Educational Psychology Review. It analyzed 74 studies on reading-based language learning and found positive effects across all seven measured domains:

The analysis also found that effects were larger when learners had some form of accountability (reading logs, brief quizzes) and when text selection was guided rather than fully open.

An earlier analysis by Nakanishi (2015) examined 34 studies with 3,942 learners and reached the same conclusion: regular reading reliably improves reading ability.

How much vocabulary do you actually learn?

This is where the numbers get specific.

Webb, Uchihara, and Yanagisawa (2023) analyzed 24 studies on vocabulary learning through reading — picking up words naturally from context, without flashcards or word lists. Their findings:

But the retention rates tell a different story:

You learn roughly 9–18% of the new words you encounter through reading. That sounds low, but it compounds: if you encounter 20 new words per reading session, you learn 2–4 of them per session. Over months, that adds up.

The comparison with flashcards is important: explicit vocabulary study produces about 39% retention on delayed meaning-recall tests (Webb, Yanagisawa & Uchihara, 2020) — roughly 3x higher than reading alone. But flashcards do not build fluency, comprehension, or reading speed. The two approaches serve different purposes.

What is the 95% rule for reading?

Schmitt, Jiang, and Grabe (2011) studied 661 learners across 8 countries and found that reading comprehension depends directly on how many words you already know in the text.

A later replication by Kremmel et al. (2023) found high individual variation at every coverage level — some learners comprehend well at 95%, others struggle even at 98%. The 95–98% range is a useful guideline, not a hard threshold.

For practical purposes: aim for material where you understand at least 95% of the words. 98% is even better. If you are constantly looking up words, the text is too hard.

What are the limitations of extensive reading?

No method is perfect, and the research on extensive reading has real gaps.

The benefits are smaller than often claimed. Hamada (2020) re-analyzed extensive reading studies using stricter methods that account for pre-existing differences between learner groups. When those differences are controlled for, the improvement from a semester of reading drops from "medium" to "small but real." Reading helps, but it is not a shortcut.

Vocabulary learning through reading is slow. Waring and Takaki (2003) tracked learners through a graded reader and found that three months later, they could recall the meaning of just 3.6% of the new words they had encountered. To have a 50% chance of recognizing a word after three months, learners needed to see it at least 8 times. This is why combining reading with deliberate review works better than reading alone.

The 95% rule is a guideline, not a law. Kremmel et al. (2023) replicated earlier research on vocabulary coverage and found that some learners comprehend well at 95%, while others struggle even at 98%. Individual factors — your ability to guess from context, your background knowledge, the type of text — matter just as much as the percentage of known words.

Most studies are in classrooms, not at home. About 59% of extensive reading studies took place in universities with structured programs. Whether the same effects hold for self-directed learners reading on their own is less well established — though the underlying mechanism (repeated exposure to language in context) should work regardless of setting.

What this means in practice

The research points to a specific combination:

1. Read daily at a comfortable level. The 95–98% coverage zone is where learning happens most efficiently. If it feels hard, go easier.

2. Volume matters more than difficulty. In one study (Suk, 2017), swapping just 30 minutes per week from grammar drills to free reading nearly quadrupled vocabulary gains and doubled reading speed improvement. Consistent, moderate reading beats occasional hard reading.

3. Review key words separately. Reading alone produces broad exposure but slow retention. If you want to lock in specific vocabulary, review it — flashcards, writing, conversation.

4. Stick with it. The research consistently shows that effects grow with time. A week of reading does little. Months of daily reading compounds into real, noticeable improvement.

5. Let something pick your level for you. The Sangers (2025) analysis found that learners improved more when someone — a teacher, an app, an algorithm — matched text difficulty to their level, rather than leaving the choice entirely open.

ToTo matches a new story to your CEFR level every day — calibrated so you understand about 95% of the words. Three minutes. Daily. Free on Android. See the research behind ToTo's approach.