Quick answer: Listen and Repeat is the first TOEFL 2026 Speaking task: 7 items, no preparation time, machine-scored against predefined answers. You hear a sentence — it is never shown on screen — and repeat it exactly. No template can help here, because you are not composing anything. What raises your score is chunking (holding meaning units instead of individual words), shadowing practice, and deliberate span extension.
This is Part 1 of the TOEFL 2026 Speaking Frameworks series.
What the task actually looks like
- A picture stays on screen throughout the task.
- The sentences you hear relate to that picture. Arrows, highlighting, or shapes may be added to direct your attention.
- Sentences get progressively longer and more complex across the 7 items.
- Recording time is roughly 8–12 seconds per item.
- There is no preparation time and no second listening.
Why templates cannot apply
A template is a reusable structure for content you generate yourself. Listen and Repeat gives you the content. There is nothing to structure, nothing to fill in, nothing to pre-plan.
What is being measured is narrower and more mechanical:
| Skill | What it means here |
|---|---|
| Phonological working memory | How much spoken language you can hold accurately for a few seconds |
| Listening precision | Catching unstressed function words (a, of, has been) |
| Pronunciation accuracy | Producing what you heard intelligibly |
| Fluency under pressure | Delivering it without hesitation or restart |
These are trainable — but through drilling, not memorisation.
Technique 1 — Chunk, don't word-count
Weak repeaters try to store a sentence as a list of words. Working memory holds roughly a handful of items, so an 18-word sentence overflows immediately.
Strong repeaters store meaning units. The same sentence becomes 4 items instead of 18:
"The students in the back row / were asked to move / to the front of the hall / before the lecture began."
Drill: take a long sentence, mark the chunk boundaries with slashes, read it aloud chunk by chunk, then repeat it from memory. Chunk boundaries usually fall at prepositional phrases, clause starts, and after subjects.
Technique 2 — Shadowing
Shadowing means repeating audio as it plays, one or two words behind, without waiting for it to finish.
Daily 10-minute routine:
- Pick 60 seconds of clear spoken English (a lecture clip, an interview, a news segment).
- Play it and speak along, trailing slightly.
- Repeat the same clip three times. Fluency improves within one session.
- Record the third attempt and listen back once.
Shadowing trains ear and mouth together. It is the single highest-return drill for this task, and it also improves your Listening band.
Technique 3 — Extend your span deliberately
The task ramps up in difficulty, so your practice should too.
| Week | Sentence length | Target |
|---|---|---|
| 1 | 8–10 words | Perfect accuracy, no hesitation |
| 2 | 11–14 words | Accuracy with one chunk boundary |
| 3 | 15–18 words | Two or three chunk boundaries |
| 4 | 18–22 words | Complex clauses, embedded phrases |
Do not jump to long sentences early. Accuracy at a shorter length is what builds the memory capacity for the longer ones.
Technique 4 — Do not self-correct
This is the most common avoidable error. You mis-say a word, stop, and start the sentence again — and the recording window closes before you finish.
Rule: keep going. A single imperfect word inside a complete, fluent sentence costs less than a restart that leaves the sentence unfinished.
Technique 5 — Use the picture
The image on screen is not decoration. It provides semantic context that supports recall — if the sentence describes something visible, the picture gives your memory an anchor.
Drill: practise with a photo open in front of you. Have a partner or a text-to-speech tool read sentences about the image, and repeat them. Train yourself to glance at the visual rather than stare at the screen blankly.
What English learners typically lose points on
| Problem | Fix |
|---|---|
| Dropping word endings (-s, -ed, -'s) | Slow shadowing focused only on endings |
| Missing unstressed function words | Transcribe short clips by hand; you will see what you skip |
| Freezing on an unknown word | Say your best approximation and continue — silence scores nothing |
| Rushing and slurring | Match the original tempo, do not exceed it |
| Long sentences collapsing at the end | Chunk practice, working backwards from the final chunk |
A 10-minute daily drill
| Minutes | Activity |
|---|---|
| 0–3 | Shadow one 60-second clip, three passes |
| 3–7 | 10 repeat items at your current target length |
| 7–9 | Re-do the 3 items you got wrong |
| 9–10 | Record one item, listen back, note one specific fix |
Consistency matters far more than session length. Speech production is a motor skill — daily short practice beats weekly long practice.
Frequently asked questions
How many Listen and Repeat items are on the TOEFL 2026?
Seven items, forming the first of the two Speaking tasks.
Is the sentence shown on screen in Listen and Repeat?
No. You hear the sentence and must repeat it from memory. A related picture stays on screen throughout.
Can I use a template for Listen and Repeat?
No. The task supplies the content, so there is nothing to structure in advance. Preparation should focus on chunking, shadowing, and pronunciation accuracy.
How is Listen and Repeat scored?
It is machine-scored against predefined answers, per ETS's 2026 test blueprint.
What should I do if I forget the middle of the sentence?
Say what you have, fluently, and keep the sentence structure intact. Stopping to recover generally costs more than the missing words.
Next in this series: Opinion questions in Take an Interview.
Verify current specifications on the official TOEFL site. Last verified: July 2026.
Put these strategies to work with daily TOEFL practice and instant AI feedback.
Start for free