IELTS Listening Part 3 is a conversation between two main speakers in an educational or training context — typically two students discussing an assignment, or a student and a tutor reviewing work. Ten marks, and the hardest part of the section for most candidates.
The difficulty is not vocabulary. It is attribution: the speakers disagree, change their minds, and half-agree, and the questions ask who thought what.
|
Want to FINALLY Pass the IELTS Academic Test And Reach Your Goals? Get Your FREE Lume Account Today to Unlock… ✅ 10,000+ PRACTICE QUESTIONS with detailed answers ✅ FREE MOCK TESTS with fast & accurate AI scoring ✅ Get BETTER SCORES with a PERSONALIZED study plan (Limited seats available!) |
| Feature | Detail |
|---|---|
| Speakers | Two main speakers, sometimes a third |
| Context | Educational or training — assignments, research, course choices |
| Questions | 10 |
| Common task types | Multiple choice with long options, matching, sentence completion, flow-chart completion |
| Main difficulty | Attributing opinions to the right speaker |
| Heard how many times | Once |
Questions 1-10.
|
Transcript Tutor: Right, Nadia, Joachim — where are you with the urban heat project? Nadia: We have the temperature data for all twelve sites, but the two rooftop sensors failed in July, so those two have a three-week gap. Joachim: I wanted to drop those two sites entirely. A gap in the hottest month is exactly the period the whole study is about. Nadia: And I said we should keep them and flag the gap. If we drop them we lose both of our rooftop locations, and then we cannot say anything about roof surfaces at all. Tutor: Nadia is right on that point. Ten sites with a caveat is a stronger dataset than eight sites with a hole in the design. Flag it in the methods section. Joachim: Fine. Though it does mean the July comparison is only across ten sites. Tutor: It does, and you should say so. What about the survey? Nadia: Two hundred and forty responses, which is more than we expected. But the age distribution is skewed — almost nobody under twenty-five. Joachim: I think that is because we distributed it through the residents association. Their membership is mostly older homeowners. Nadia: That was my assumption too at first, but I checked the timestamps. The under-twenty-fives who did respond nearly all came in through the one social media post, and that post only ran for two days before it was taken down. Joachim: So it is a distribution channel problem rather than a membership problem. Nadia: Right. Which is fixable — we could run the post again for a fortnight. Tutor: Do that, but be careful. If you extend collection you will have two waves gathered under different weather conditions, and perception of heat is weather-dependent. You will need to report them separately. Joachim: Could we not just weight the sample instead? Tutor: You could, and many published studies do, but weighting a group you have almost none of does not fix the problem — it amplifies whatever those few people said. Collect more. Nadia: Then we should start the second wave immediately, so the weather difference is as small as possible. Tutor: Yes. And one more thing — your literature review is still built almost entirely on studies from temperate cities. There is a substantial body of work from subtropical contexts, and a reader will ask why you ignored it. Joachim: I did find those but left them out because the climate is not comparable. Tutor: That is an argument to make explicitly in the review, not a reason to omit them. Say what you excluded and why. Otherwise it reads as though you did not look. |
Questions 1-5: Choose the correct letter, A, B or C.
Questions 6-10: Who expresses each view? Write N for Nadia, J for Joachim, or T for the tutor.
| View | Your answer |
|---|---|
| 6 Losing both rooftop sites would remove a whole surface type from the study | |
| 7 The two collection waves must be reported separately | |
| 8 The problem lies in how the survey was distributed, not who the members are | |
| 9 The second wave should begin as soon as possible | |
| 10 Excluded literature should be named and justified rather than omitted |
| Q | Answer | Evidence | Trap |
|---|---|---|---|
| 1 | B | “the two rooftop sensors failed in July… a three-week gap” | C over-generalises to all twelve sites |
| 2 | A | “I wanted to drop those two sites entirely” | This is his original position, which he later abandons |
| 3 | B | “Ten sites with a caveat is a stronger dataset… Flag it in the methods” | A is Joachim’s view, not the tutor’s |
| 4 | B | “I checked the timestamps… came in through the one social media post” | Nadia says she first agreed with Joachim, then changed her mind |
| 5 | B | “it amplifies whatever those few people said” | A is contradicted: “many published studies do” |
| 6 | N | “then we cannot say anything about roof surfaces at all” | The tutor agrees with it, but Nadia says it |
| 7 | T | “You will need to report them separately” | — |
| 8 | J | “So it is a distribution channel problem rather than a membership problem” | Nadia supplied the evidence; Joachim states the conclusion |
| 9 | N | “we should start the second wave immediately” | The tutor says “Yes” — agreement is not authorship |
| 10 | T | “Say what you excluded and why” | — |
Questions 6, 8 and 9 are the pattern that decides Part 3 scores. In each, one speaker states the view and another agrees with it. The question asks who expresses it, and agreeing is not expressing.
| Technique | How |
|---|---|
| Initials, not names | Write N, J, T beside each option as you hear it — there is no time for words |
| Mark changes of mind | An arrow from the old position to the new one; the new one is almost always the answer |
| Separate evidence from conclusion | The person who supplies a fact is often not the person who draws the conclusion from it |
| Note agreement separately | A tick means “agreed with”, not “said” |
| Listen for the tutor’s verdict | When a tutor adjudicates a student disagreement, that verdict is usually a question |
| You hear | It means | Do |
|---|---|---|
| “That was my assumption too at first, but…” | A change of mind is coming | Prepare to overwrite |
| “Fine. Though it does mean…” | Reluctant concession with a reservation | Both parts may be tested |
| “You could, and many studies do, but…” | Partial concession, then rejection | The answer is after the “but” |
| “X is right on that point” | Adjudication | Attribute the view to X, not to the adjudicator |
| “So it is A rather than B” | A conclusion being formulated | Note who says it |
| “That is an argument to make, not a reason to…” | Reframing, not agreement | Do not record it as agreement |
| Factor | Effect |
|---|---|
| Two or three speakers, overlapping views | Answers depend on who said it, not just what was said |
| Academic register and abstract nouns | Less concrete language to anchor on |
| Long multiple-choice options | More reading during the audio |
| Changes of mind built into the script | The first plausible answer is often the trap |
| No numbers or spelled names to catch | Nothing easy to bank early in the part |
A conversation between two main speakers in an educational or training context — usually students discussing an assignment or a student in a tutorial. It carries 10 of the 40 Listening marks, and a third speaker such as a tutor may take part.
Because the questions test attribution as much as comprehension. Speakers disagree, concede and change their minds, and the options are built from views that all appear in the recording, so following the meaning is not enough.
Write initials next to the options as you hear them rather than words, and mark a change of mind with an arrow. When one speaker states a view and another agrees, record it under the speaker who stated it — that is what the question asks.
Yes, and it matters more in Part 3 than anywhere else, because the options are long. Underline the single word that differs between them so you can choose without re-reading while the audio runs.
Leave it and move forward. Questions follow the order of the recording, so a missed answer is behind you and the next one is being spoken now. Chasing one lost mark routinely costs two or three.
No. Every part of IELTS Listening is heard once only.
|
Want to FINALLY Pass the IELTS Academic Test And Reach Your Goals? Get Your FREE Lume Account Today to Unlock… ✅ 10,000+ PRACTICE QUESTIONS with detailed answers ✅ FREE MOCK TESTS with fast & accurate AI scoring ✅ Get BETTER SCORES with a PERSONALIZED study plan (Limited seats available!) |
Share
Lume Test
Editor
Lume Test is a leading platform offering practice questions and mock tests for major English exams like PTE, IELTS, TOEIC, Duolingo English Test, and more. We provide up-to-date, comprehensive materials designed to help learners improve their language skills and achieve high scores. With expert-crafted tests that closely mirror real exam formats, Lume Test ensures you're fully prepared for success. Your goals are our mission!
