Dictation for Language Learning: How and Why It Works

Dictation is a listening exercise. You hear a short passage, write down what you heard, then compare your version with the original and fix the gaps. The one controlled trial we found reports a gain in listening, and a teaching guide from Hong Kong's Education Bureau says dictation also builds phonics and writing skills ¹. You can run it alone with any audio, in short pieces, or in class with a partner. Our overview of dictation languages and accents covers the other ways speech tools handle the languages you are learning.

Dictation trains listening, spelling and grammar in one exercise

Dictation makes you do three jobs at once. To write a passage you have to hear each word, decide how it is spelled, and get the grammar right, including endings and word order you did not catch on the first listen. Reading a text or answering questions about it lets you skip the first job. Dictation does not let you skip it.

That is our explanation of why the exercise helps, not a finding from a study. The research below measures outcomes such as listening scores and learners' views. None of it measures the mechanism directly.

Teachers have used dictation as a test for a long time. Cambridge's dictionary describes a school dictation as a test in which a passage is read to students learning a language, and they write it down ². Its business dictionary also defines dictation as speaking aloud for a person or a machine to record ³. The learner-practice sense is the one this guide covers. The Hong Kong handbook says dictation is more than a testing device, and that it helps teachers see how learners are progressing ⁴.

One 2002 trial supports dictation for listening, and no other study we found measured it

The strongest result comes from a 2002 controlled study of elementary EFL learners at a language institute in Tehran, Iran ⁵. Both groups used the textbook's listening exercises. One group also had dictation, 11 times in a single term ⁶. The term was 20 sessions long ⁷, so the 11 dictations equal 55 percent of the sessions, if each one fell in a separate session ⁸. The authors report that dictation had a significant effect on listening comprehension in the experimental group ⁵.

That single trial is a useful signal, not proof. It ran at one institute over one term, and the abstract reports significance without saying how large the gain was. A second study, by Muriel Gallego in 2014, tested a close relative of dictation called dictogloss with 497 learners at two levels ⁹. It measured how learners felt about the task, not how much they improved ¹⁰. The table sets out what each study measured.

Two gaps matter for a learner. None of these studies measured speaking, and none measured spelling or grammar directly. The case that dictation builds those skills rests on the handbook guidance above and on how the exercise works, not on a controlled measurement. If you want the strongest claim you can make, it is this: dictation improved listening in one controlled trial, and it is reasonable to expect spelling and grammar to improve with it.

The method: listen, write, compare, repeat

You need only audio and a text to check against. The routine below works with any language you are learning.

1. Pick a short passage of one or two sentences, from a lesson, a learner podcast or a short clip you can replay.
2. Play one segment, a sentence or clause, and write it down without looking back at the audio.
3. Replay the segment as often as you need, and write the words you still missed. Mark them so you can see where you struggled.
4. Compare with the original text. Label each error by type: a spelling error, a missing word, a wrong ending or tense, or a homophone such as "their" for "there".
5. Repeat the passage a day or two later until your first attempt has no errors.

Your error labels from step 4 tell you what to practice next. The readings below are a rule of thumb from how dictation works, not a validated scale.

If you only have audio, you can make a rough transcript with our free audio to text tool in the browser. Correct it by ear before you use it, because a recognizer can mishear words and you would then practice the wrong spelling. The step-by-step order above is advice rather than a tested protocol. It is what the sources we read describe, and no study we found compared it with other routines.

Choose audio slightly above your current level

Stephen Krashen's input hypothesis holds that learners move forward when they understand input slightly more advanced than their current level, which he called i+1 ¹¹. Dictation applies the same idea in a stricter form: the audio must be clear enough to hear and hard enough to make you work.

A practical check follows from that idea. If you miss most of the words on the first listen, the passage is too hard, so choose something slower or more familiar. If you write it perfectly the first time, it is too easy. The exact threshold is our rule of thumb, not a measured figure.

Choose recordings with the accent you will actually hear. A single speaker trains you for one accent, not for every speaker of the language, and an accent can change vowels and stress enough to change which words you hear. The non-native speech of other learners is a different kind of input, and it is worth adding once you are comfortable with one accent. Krashen's i+1 is a theoretical idea, not a measured quantity, so use the check, not a number, to pick your audio. For a daily routine, see our question on how to practice dictation in English.

Partial dictation, dictogloss and running dictation add variety

Full dictation is the core exercise. Three variants change the task so that it is easier, more social, or more focused on a sound or structure. The table shows what each asks of the learner and what the evidence says.

Running dictation is the most physical of the four. The writer sits at a desk, the runner walks or runs to the posted text, and they swap roles partway through. The Language Teacher's running dictation guide asks the teacher to remind students to focus on their pronunciation, which makes it a useful bridge to speaking practice. It reports no measured results, so treat it as a procedure, not as evidence.

Dictogloss asks learners to reconstruct a short text they heard, and the work is done with a group. Gallego's study found the technique better received by advanced-low learners, and most learners found it useful ¹⁰. Learners' views are not gains, so do not treat that result as proof that dictogloss improves grammar.

Voice typing turns dictation around to check your pronunciation

Most dictation guides describe the exercise in one direction: you hear, you write. Voice typing works the other way. You speak a sentence, and the phone or computer writes what it heard. If the output differs from what you meant, that mismatch is a clue about your pronunciation, your stress, or the recognizer's own error. Both sides can be wrong, so treat a mismatch as a question to check, not a verdict.

The best evidence for this use comes from a controlled study of an automatic pronunciation trainer. It was built on a phoneme-based speech recognition model and gave explicit corrective feedback on Arabic phrases. Twenty-six American college students took part ¹³. The trainer group performed significantly better on the posttest than the in-class control group ¹⁴. That is a dedicated teaching tool with explicit feedback, not a general dictation app, and it covered three phrases ¹⁴. It shows that recognition-based feedback can help, not that any voice typing will.

Accent is the main limit. A 2024 analysis of OpenAI's Whisper speech model found better performance on North American English accents than on British and Irish English accents ¹⁵. A recognizer can therefore mishear a pronunciation that is correct for a different English. Research on foreign-accented speech suggests the problem can be reduced. In a 2007 study, the best combination of pronunciation modelling and acoustic adaptation cut word error rate by 46 to 71 percent relative ¹⁶. Those are research systems, not the dictation built into your phone, so read the mismatch with that in mind.

Here is a routine for checking your pronunciation with voice typing:

1. Set one language to the one you are learning. On a Mac running macOS 27, Apple's Mac Help describes going to Keyboard settings, opening Dictation, clicking the Edit button next to Languages, and selecting a language and region ¹⁷. Apple says you can set up Dictation for several languages and switch between them as needed ¹⁸, but its page says nothing about speech that mixes two languages in one sentence, so set one language at a time while you practice.
2. Speak one sentence slowly, as you would to a listener, and read the output.
3. Mark each word that differs from what you said.
4. Change one sound in a marked word, such as a vowel or a stressed syllable, and say it again to see whether the output changes.
5. Check the word in a dictionary with audio. If the dictionary's pronunciation matches yours and the recognizer still rejects it, the error is probably the recognizer's.

On iPhone, dictation may not be available for every language or region ¹⁹, so confirm your language is listed before you build a habit around it. For readers who want the wider picture of how these errors are measured, see our guide to speech recognition accents and our overview of word error rate. Our question page on voice typing for pronunciation covers the same routine in more detail.

Yap, which we make, runs its dictation on the device on Mac and Windows. For this kind of practice the built-in tools above are enough. Our dictation in the browser also works for a quick check.

The evidence covers listening and learners' views, not speaking

Put the research together and the picture is narrower than the claims made for dictation. The controlled trial measured listening ⁵. The dictogloss study measured learners' views ¹⁰. The pronunciation trial used a recognition tool with explicit feedback, not dictation itself ¹⁴. No source we reviewed measured speaking after dictation, or whether gains last after a term.

That gap matters for planning. Use dictation to strengthen listening and spelling, and use speaking practice, such as the running dictation routine or the voice typing check, to work on pronunciation. Our guide to how to get used to dictation covers the habit side. To see which languages each device supports, read our guide to dictation languages.

At school, a dictation test checks spelling from spoken words

In a classroom, dictation is usually a teacher-led test of spelling and listening. The student writes, and the teacher reads. That is the sense the Cambridge dictionary gives for schools ², and the Hong Kong handbook describes it as a classroom activity for phonics and writing ¹. Voice typing is different: the machine writes what you say, and the student does not need to spell.

Parents and teachers often search for word lists, and we answer those questions on separate pages: English dictation words for young children, Hindi dictation words without matra, and what dictation is called in Hindi. For multilingual learners, the guides on multilingual dictation and Hindi voice typing cover the device side. Our general question on whether dictation is good for language learning gives a shorter answer with the same sources.