What tasks are in the Listening section?
ETS names four task types in Listening: “Listen and Choose a Response,” “Listen to a Conversation,” “Listen to an Announcement,” and “Listen to an Academic Talk.” Its content page lists 47 questions for the section and says the test is adaptive, so duration and question count may vary; 47 is the figure in ETS’s table, not a guarantee for your test.
ETS’s Listening page says the section measures the ability to understand spoken English in modern learning environments through short recordings focused on meaning, organization, and speaker intent. The table below summarizes its descriptions of each task and adds the question counts attributed to each in ETS’s 2026 “Blueprint and Specifications.” Written for the 2026 update, the document warned that minor revisions were possible before launch; treat its ranges as estimates.
These ranges do not add up to one fixed total, and the reviewed documents do not say which combination a test taker receives.
| Task, ETS name | What the task requires, according to ETS | Expected questions |
|---|---|---|
| Listen and Choose a Response | Listen to a short utterance and choose the most appropriate response: meaning, intent, and common spoken-English expressions in an academic setting. | 15 to 19 |
| Listen to a Conversation | Follow a conversation, usually about campus life: main ideas, key details, speaker intent, and implied meaning. | 10 |
| Listen to an Announcement | Understand an announcement in a university setting: speaker purpose, key information, implied meaning, and likely next actions. | 6 to 10 |
| Listen to an Academic Talk | Follow a short talk by a professor or expert: main ideas, details, organization, inferences, and sometimes uncommon or idiomatic vocabulary. | 8 to 16 |
Does Listening adapt to your level?
Yes, in two stages. ETS’s specifications document says Reading and Listening are designed as two-stage adaptive tests, while Writing and Speaking are linear: all test takers on the same form receive the same tasks. For Listening, its table distinguishes a first module, the “Router,” followed by a “Lower” or “Upper” module. ETS’s page for advisors describes this adaptive technology as a way to personalize the testing experience.
The same table estimates durations in minutes: 18 for the Router, then 7 for the Lower module or 11 for the Upper module. ETS says the total includes the Router plus one of those modules. The “Approx. Base Time” of 29 on the content page equals 18 + 11; that is our calculation, not an ETS statement, and we do not infer a test duration from it. ETS advises allowing about two hours for the full test and says duration varies.
The pages reviewed do not say how ETS routes test takers to the lower or upper module. Practical advice, not an ETS rule: do not try to guess which module you are in; answer every question as if it counts.
How many times do you hear the audio, and can you take notes?
Once. ETS’s specifications document says recordings are played once and accompanied by a visual representation of the speaker or speakers. It describes short utterances of up to six stressed syllables and talks of up to 250 words; intermediate recordings are usually 35 to 100 words, at a natural speaking rate. Voices are AI-generated and sometimes recorded by actors, with a balance of accents that the document says it aims for; the Listening page lists native accents from North America, the United Kingdom, New Zealand, or Australia. After each recording, the document describes questions usually presented with a printed prompt and multiple-choice options, individually or in sets depending on recording length.
Notes are allowed: the Listening page says you may take notes while listening to answer questions. Materials depend on location. At a test center, the administrator provides pencils and scratch paper, and the Bulletin prohibits bringing your own; you can request more only after using and returning all pages already received. At home, the Bulletin allows a small desktop whiteboard with an erasable marker or a sheet of white paper inside a transparent sleeve with an erasable marker; the paper must never leave the sleeve.
The Bulletin also prohibits using scratch paper before the test, between sections, or during breaks, and says no breaks are scheduled. During the section, take notes without losing track of the audio; this is our strategy advice.
How is Listening scored from 1 to 6?
ETS reports each section on a 1–6 scale in half-point increments, and the overall score is the average of the four sections, rounded to the nearest half band. For Listening, the specifications document lists a maximum of 1 point per scored question and counts 47 questions and 35 raw points in its table. It also says the section includes unscored questions used to collect data for future test versions; it does not give their number or say how to identify them, so treat every question as scored.
The reviewed ETS documents do not publish a table converting raw points to a 1–6 band. Its April 2026 report describes how the 1–6 scale is linked to the former 0–30 scale and the CEFR, but gives no Listening raw-score conversion. The sources reviewed therefore do not support a “number of correct answers needed for a 5.”
At the end of the test, you see unofficial Listening and Reading scores; the Bulletin warns they may differ from official scores because of post-test statistical analysis. To receive scores, you must have answered at least one Listening question, in addition to the minimum responses required in the other three sections. The next table summarizes ETS’s Listening descriptors for each CEFR level in the appendix to its April 2026 report.
These descriptors explain what a score means; they do not set admissions cutoffs. ETS leaves each institution to set its own requirement on the scale it publishes. Do not convert a figure from one scale to another.
| Listening score | CEFR | What ETS describes |
|---|---|---|
| 6 | C2 | Make appropriate inferences when connections or implications are not explicit. |
| 5 – 5.5 | C1 | Infer attitude, mood, and intent from contextual, grammatical, and lexical clues and anticipate what comes next; follow most lectures and discussions with ease. |
| 4 – 4.5 | B2 | Follow extended speech and complex arguments when the topic is sufficiently familiar and the argument is guided by explicit markers. |
| 3 – 3.5 | B1 | Understand the main points of clear, standard speech on familiar topics. |
| 2 – 2.5 | A2 | Understand the gist of short recordings on predictable everyday topics. |
| 1 – 1.5 | A1 | Identify specific information, such as places and times, on familiar topics when spoken slowly and clearly. |
What strategy works for each task?
The following is our strategy, based on what ETS says it measures; these are not ETS rules. The idea is that because the recording is not replayed, listen for the function and meaning of what is said, not just the words.
The three exercises below are written teaching examples. The conversations are fictional; the talk simplifies the bucket-measurement problem described by the Met Office, cited as a source. There is no audio here: each prompt includes a transcript, but on test day you hear it once without reading it. To approximate the test, read the transcript once from beginning to end without going back, then answer.
- Listen and Choose a Response: identify the utterance’s function—a request, offer, complaint, or question—then choose a response that addresses it. Eliminate choices that repeat a word without responding or answer a different question.
- Listen to a Conversation: identify who is speaking, what each person wants, and what is decided. A seemingly off-topic sentence is often an indirect response; ask yourself what the speaker is trying to accomplish by saying it.
- Listen to an Announcement: note the purpose (announce a change, remind, request), key information (date, location, instruction), and what listeners will probably do next.
- Listen to an Academic Talk: follow the structure rather than individual sentences—a claim, example, objection, and conclusion. Infer unfamiliar words from context; ETS says talks sometimes include uncommon or idiomatic vocabulary.
- Notes: keywords, numbers, arrows—never full sentences.
Transcript Speaker: “Could we push the lab report deadline to Friday?” Choose the most appropriate response.
Choose an answer.
The solution, step by step
A: the utterance is a request, and only answer A responds to it.
The method in action
- First, identify the utterance’s function. “Could we push… to Friday?” asks to postpone something. A good response accepts, declines, or suggests an alternative; A agrees to the extension with a limit, so it is a complete response.
- B repeats the sound of “push” with a different meaning, as in pushing a door; this word-reuse trap exploits listening for words instead of meaning. C repeats “report” and “Friday” without addressing the request. D begins with “Yes,” which may sound reassuring, but answers a question about the lab’s location that no one asked.
Transcript Maya: Did you finish the poster for Thursday? Leo: Almost. The charts are done, but the text is a mess. Honestly, I’m thinking of cutting the methods box altogether. Maya: Hmm. Professor Duarte did say the methods count for a third of the mark. Leo: I know, but there’s no room. Everything’s too small to read from a metre away. Maya: What if you shrink the charts instead? People can read the figures in the captions. Leo: That might work. I’ll try it tonight and send you a photo. Why does Maya mention Professor Duarte?
Choose an answer.
The solution, step by step
A: Maya responds to the idea of removing the methods box by reminding Leo what the professor said about it.
The method in action
- The question asks about intent, not a detail. Leo suggests removing the methods box; Maya counters with the fact that methods account for part of the grade. She does not say “keep it” explicitly; the meaning is implied, which is what the task measures.
- B is incorrect: Thursday is the only deadline mentioned, and no one suggests moving it. C is incorrect: Leo says everything is too small, and the professor is not connected to that issue. D is incorrect: Leo says the text is “a mess,” and Maya does not disagree.
Transcript Historical sea surface temperature records include measurements taken from ships using buckets. A sailor would lower a bucket over the side, haul it up, read a thermometer, and write the number in the logbook next to the date and the position. Climate scientists now collect those logbooks. But there is a catch. A wooden bucket cools slowly, while a canvas bucket, hauled up on deck, loses heat to the wind. So the same sea gives different readings, depending on the bucket. Researchers can’t simply throw the old numbers away. Instead, they use records and other evidence to identify or estimate how the measurements were taken, then adjust the data for those methods. What is the speaker’s main point?
Choose an answer.
The solution, step by step
A: the main point is that old measurements remain useful once adjusted for the type of bucket used.
The method in action
- Follow the talk’s structure: context (old ship measurements), a resource (logbooks), a problem (the bucket changes the reading), then a solution (identify or estimate the measurement method and adjust the data). The main idea is this solution to the problem.
- B claims something the talk never says: it does not say who chose the bucket; bucket type is the problem, not the main point. C is incorrect: the talk does not compare how regularly sailors measured temperatures with satellites. D reverses the talk: researchers “can’t simply throw the old numbers away”; they adjust the records instead of discarding them.
The pitfall
Should you write down everything you hear?
A common tip is to write as much as possible so you have “everything” when you answer. With a recording played once at a natural pace and potentially reaching 250 words, this can backfire: writing one sentence may cause you to miss the next. ETS describes questions about meaning, intent, organization, and inference—things you cannot simply copy down. ETS does not say what to note; our advice, which you can test with your own recordings, is that keywords, numbers, and logical connections work better than sentences.
A second tip to avoid is practicing with pre-2026 Listening materials without checking the tasks. ETS says its January 2026 update introduced new task types, and the TOEFL iBT guide recommends checking that a book or study plan covers current tasks. Compare your materials with the four task names above.
Key takeaways
- ETS lists four task types: Listen and Choose a Response, Listen to a Conversation, Listen to an Announcement, and Listen to an Academic Talk; its table lists 47 questions, but the number may vary by test.
- According to ETS’s specifications, each recording plays once; notes are allowed on the paper or whiteboard specified in the Bulletin.
- Reading and Listening use two-stage adaptive testing; the reviewed ETS documents do not give the routing rule, so answer every question as if it counts.
- Scores range from 1 to 6 in half-point increments; the reviewed ETS documents do not map correct answers to scores, and institutions’ requirements must be read on the scale they publish.