Listening section: format, scoring, and what the adaptive second module changes
The updated TOEFL iBT Listening section is multistage adaptive: you first do a routing module, then ETS routes you to an easier or harder second module based on that performance. The whole section is about 29 minutes with 47 items, but the exact mix and timing can vary slightly.
| Task | Typical total items | What you do |
|---|---|---|
| Listen and Choose a Response | 15–19 | Hear one utterance, choose the best printed reply |
| Listen to a Conversation | 10 | Hear a short two-speaker conversation, answer 2 questions |
| Listen to an Announcement | 6–10 | Hear a short announcement, answer usually 2 questions |
| Listen to an Academic Talk | 8–16 | Hear a short academic talk, answer 4 questions |
Scoring facts:
- Each selected-response item is worth 1 point.
- Some items are unscored pretest items and are not identified.
- ETS does not publish the raw-to-band conversion.
- On Courselo fixed-form mocks, a realistic estimate is:
| Listening share correct | Estimated band |
|---|---|
| 92%+ | 6 |
| 85%+ | 5.5 |
| 75%+ | 5 |
| 66%+ | 4.5 |
| 56%+ | 4 |
The timing plan that actually works
ETS does not publish the exact timing of each adaptive module, only approximate section totals and task ranges. Use this conservative pacing plan.
Section-wide pacing
| Task | Target time per item/set | What to do |
|---|---|---|
| Listen and Choose a Response | 8–12 seconds after audio | Decide by speech act, do not overthink |
| Conversation questions | 20–25 seconds each | Use notes only to confirm the answer |
| Announcement questions | 20–25 seconds each | Find purpose + action required |
| Academic Talk questions | 25–35 seconds each | Use structure notes, not full sentences |
Checkpoints inside the section
- First third of the section: maximize accuracy, because this is likely to contain much of the routing module.
- Middle: protect time; no item is worth extra points, so do not spend 40 seconds proving one answer.
- Final third: fatigue rises; keep note-taking minimal and mechanical.
Task 1: Listen and Choose a Response
This is the most speed-based task. You hear one utterance and choose the most appropriate printed reply.
Fastest reliable method
Your cue is the question type or speech act:
- request
- offer
- suggestion
- apology
- complaint
- indirect question
- opinion check
- confirmation
Ask: What kind of reply is required? Not: What words did I hear?
Examples of what you should predict:
- If the speaker says, “Could you send me the notes?” → reply must accept, refuse, or ask for clarification.
- If the speaker says, “Do you know when the lab opens?” → reply must give time information, not location.
- If the speaker says, “I can’t believe the printer jammed again.” → reply should show reaction/help, not define “jam.”
Distractor patterns to reject
| Distractor pattern | What it does |
|---|---|
| Echo | Repeats a word from the audio but does not answer the function |
| Wrong wh-type | Answers where instead of when, who instead of why |
| Wrong tense/person | Grammatically possible but mismatched to the situation |
| Literal/idiom trap | Interprets an idiom word by word |
Rules
- Do not take notes. They slow you down.
- Decide in one pass.
- If two options seem possible, choose the one that fits the social function better.
Task 2: Listen to a Conversation
You hear a short two-speaker exchange and answer 2 questions. The questions usually test gist/purpose, detail, implied meaning, or prediction.
What to note
Write only this frame:
- Topic
- Problem/change
- Decision/next step
Example note style:
club poster
room unavailable
move to library / email members
That is enough for almost every 2-question set.
Fastest reliable method
- In the first lines, identify why they are talking.
- Listen for the turning point: but, actually, instead, I thought, turns out.
- At the end, catch the action: what one speaker will probably do next.
Common question cues
- Main purpose: answer from the whole conversation, not one detail.
- Implied meaning: focus on tone + context, especially after hesitation, correction, or contrast.
- Prediction: the answer is usually the final action or agreed plan.
Task 3: Listen to an Announcement
Announcements are short and dense. Most wrong answers miss either the purpose or the required action.
What to note
Use this 4-part template:
- Who is speaking / context
- What changed
- Who is affected
- What listeners should do
Example:
orientation office
tour delayed 30 min
new students
wait in hall / check text update
Fastest reliable method
Announcements often follow a fixed order:
- purpose
- key detail
- instruction
So if you lose one detail, keep listening for the instruction. That often answers one question directly.
Trap patterns
- true detail from the audio, but not the main purpose
- old plan instead of the changed plan
- option for the wrong group of listeners
Task 4: Listen to an Academic Talk
This is where many band-6 attempts fail. The talk is short, so the test is not about memory capacity. It is about recognizing structure.
What to note
Do not try to write content sentences. Use a skeleton:
- Topic / term
- Definition
- Example 1
- Contrast / exception
- Conclusion
Example:
biol: mimicry
one species resembles another
ex: harmless fly looks like wasp
contrast: not camouflage
benefit = avoid predators
Fastest reliable method
Listen for four signals:
- definition: “is called,” “refers to,” “means”
- example: “for instance,” “consider”
- contrast: “however,” “in contrast,” “unlike”
- function: “this matters because,” “I mention this because”
Most 4-question sets map onto those signals:
- main idea
- detail
- why mention X
- inference/organization
How to answer “Why does the speaker mention…?”
This is a function question. Your cue is not the fact itself; it is the job that fact does in the talk.
Usually the mention is there to:
- illustrate a definition
- provide evidence
- contrast with the main idea
- correct a likely misunderstanding
- make an abstract point concrete
If the speaker mentions penguins in a talk about adaptation, the answer is rarely “to give information about penguins.” It is more likely “to provide an example of how a species survives in a specific environment.”
Accent training: North American, British, Australian
ETS uses varied English accents. Your goal is not perfect accent recognition; it is stress recognition.
What changes across accents
- vowel quality
- speed of unstressed syllables
- intonation patterns
- some consonants, especially /r/ and /t/
What does not change
- discourse markers: but, so, actually, anyway, first, however
- content-word stress
- structural signals: definition, example, contrast, conclusion
Training plan
| Accent | What to train | Drill |
|---|---|---|
| North American | reduced vowels, fast linking | shadow 30-second clips |
| British | non-rhotic /r/, sharper contrast in vowels | transcribe key nouns and verbs only |
| Australian | wider vowel shifts, rising intonation | listen for stress words, not vowels |
Use a 15-minute cycle:
- 5 min listen without transcript
- 5 min check transcript and mark stress words
- 5 min shadow the clip aloud
Routing-module strategy vs harder second-module strategy
In the routing module
Priority order:
- Listen and Choose a Response: bank the fast points.
- Conversation and Announcement: avoid careless misses on purpose/action questions.
- Academic Talk: stay structured; one bad note set can cost several items.
The routing module should feel conservative. Avoid heroic guesses based on one heard word.
In the harder second module
Expect denser distractors:
- more plausible paraphrases
- closer contrasts
- more indirect function questions
Your adjustment:
- slow down after the audio, not during it
- trust structural notes
- reject options that are true in general but not true for this talk
Drills that raise Listening fastest
Highest-return drills
- Speech-act drill for Listen and Choose a Response
- 30 items per set
- label the utterance type before choosing
- 3-line notes drill for conversations
- topic / problem / next step only
- Instruction capture drill for announcements
- listen only for what changed and what to do
- Definition-example-contrast drill for talks
- one note line per function
Weekly target for a 5.5–6 candidate
| Drill | Volume |
|---|---|
| Listen and Choose a Response | 120–150 items |
| Conversations | 20 sets |
| Announcements | 12–16 sets |
| Academic Talks | 16–20 sets |
| Accent shadowing | 5 days × 15 min |
Band targets and what they mean in practice
- Band 4.5: usually good grasp of purpose and many details, but misses some implication and function questions.
- Band 5: usually accurate on everyday tasks; some losses on harder talks.
- Band 5.5: strong routing module, controlled note-taking, few avoidable errors.
- Band 6: routed to the harder second module and still highly accurate; on Courselo mocks, roughly 92%+.
Bottom line method
- Listen and Choose a Response: identify the speech act, then reject echo and wrong-wh distractors.
- Conversation: note topic, problem, next step.
- Announcement: note purpose, change, action required.
- Academic Talk: note term, definition, example, contrast.
- For “Why does the speaker mention…?” questions, answer the function, not the fact.
- In all accents, track stress words and discourse markers, not pronunciation detail.