An original learner script written for practice, marked the way an examiner-trainer would comment on it. The bands are our judgement, not an official IELTS result.
8.5 Annotated copies · Speaking
Speaking Part 3 · band 8.5
IELTS · Speaking Part 3
The task
Answer these Part 3 questions in order. Each paragraph of the fictional sample answers the corresponding question; it is a response sequence, not one uninterrupted speech.
Question 1: Why are people often persuaded by confident experts?
Question 2: Can an expert change an opinion without losing public trust?
Question 3: Who should be responsible when an organisation follows advice from a computer system?
Question 4: Should the public always be consulted about technical decisions?
Marked transcript · Band 8.5
Select a highlighted passage to see the examiner-trainer's note.
I think confidence is attractive partly because it reduces the listener's effort. If you know very little about a subject, a decisive answer gives you something to hold on to1, whereas a careful answer may leave you with several possibilities to consider. There's also a tendency to confuse being comfortable in public with knowing what you're talking about. Those qualities can go together, of course, but they don't have to2. I'd pay more attention to whether someone can explain why they're confident and what would change their view. That gives me something to evaluate beyond the performance itself.
Yes, I think an expert can change their mind without losing trust, and sometimes refusing to change would be the more worrying response. It depends on what has changed. If new evidence undermines an assumption, updating the conclusion is reasonable. If the expert is simply following whichever view is currently popular, that's different. An explanation of the assumptions help people see the difference3. I wouldn't expect an expert to predict every development, but I would expect them to say what they knew at the time and why the new information matters4. Otherwise a revision can look arbitrary even when it isn't.
I'd keep responsibility with the organisation making the decision. A computer can identify a pattern or estimate an outcome, but someone still decides what counts as a good result. Suppose a council uses a system to recommend which bus routes to retain. The system might be very accurate about passenger numbers without knowing how heavily a small group depends on a particular route5. If the council accepts the recommendation, it should explain that trade-off rather than saying the computer decided. The tool contributes to a decision; it doesn't inherit responsibility for it6. Human review needs clear reasons too, though7, or it can become an equally unaccountable judgement.
Not every technical calculation needs to be put to a public vote. I wouldn't ask residents to settle an engineering equation by majority opinion8. But the purpose of a project, the inconvenience people are expected to accept and the alternatives they value are different questions. Local knowledge can also reveal something a technical model has missed, such as when a supposedly quiet street is actually busiest. So I'd separate consultation about priorities and lived experience from checking specialist calculations, while making sure the two inform each other9. The public shouldn't be excluded just because they can't reproduce the calculation, and the calculation shouldn't be ignored just because its conclusion is unpopular10.
1
2
3→ An explanation of the assumptions helps people see the difference.
4
5
6
7
8
9
10
By criterion
Fluency and Coherence
9.0
Each written answer directly addresses its question, develops a distinction and supplies a reason or example. Topic development supports a band 9 target, but spontaneous fluency, hesitation and responsiveness to follow-ups must be assessed in a recording rather than inferred from the transcript.
Lexical Resource
9.0
The language is flexible and natural across abstract ideas: 'something to hold on to', 'undermines an assumption', 'inherit responsibility' and 'lived experience'. Technical terms are explained through ordinary examples rather than displayed for their own sake.
Grammatical Range and Accuracy
8.0
A broad range of conditionals, relative clauses, concessions and parallel contrasts is mostly accurate. 'An explanation of the assumptions help' contains one subject–verb agreement slip: the singular head explanation requires helps.
Pronunciation
8.0
Unassessed delivery target only. No audio is supplied, so the numeric field is conditional, not an observed pronunciation band. The intended delivery remains easily intelligible through the longer clauses, with thought groups and prominence clarifying contrasts between evidence, values and responsibility.
Overall8.5Atelier · IELTS
Three changes for the next band
Illustrative 8.5 target for a 417-word sequence of four answers. The organisation and lexical flexibility support high-band performance, with one isolated agreement slip to repair. FC and P remain conditional on live delivery; the transcript does not establish an actual pronunciation score or a certified overall Speaking band.
Repair the agreement after explanation, then practise several long subjects with intervening plural nouns so the same slip does not recur under pressure.
Record spontaneous answers to the four questions and add an unprepared follow-up to each; check that the qualification develops the answer rather than becoming a memorised formula.
Use audio feedback to check meaningful clause boundaries and selective prominence in the longer contrasts, preserving an intelligible natural accent rather than imitating another speaker.