Part 3 is the two-way discussion — the examiner asks 4–6 questions that extend the Part 2 topic into broader, more abstract territory. Answers should be 3–5 sentences, analytical but conversational. Use discourse markers naturally ("well," "I suppose," "having said that," "interestingly enough") and avoid sounding like a prepared speech. The examiner will probe your reasoning, so leave hooks they can follow up on.
Topics in this set:
---
Sample (Band 8.5+)
Well, I think the range is actually quite broad — it can be anything from really small, everyday favours to more significant kinds of support. On the practical side, neighbours often collect each other's post, water the plants, or keep an eye on the house when someone's away on holiday — that sort of thing. But beyond that, there's a deeper, almost informal safety-net function, you know — checking in on elderly residents, sharing spare keys in case of emergencies, or even looking after each other's children for half an hour if something comes up. I'd also say there's a less tangible but equally important dimension, which is simply being there — having someone next door who'll notice if something seems off and will knock on the door to ask if everything's alright. It's that combination of practical help and, I suppose, mutual vigilance that really makes a neighbourhood function.
---
Sample (Band 8.5+)
That's an interesting question because the answer really depends on where you look. In rural areas and older residential neighbourhoods, I'd say people still know their neighbours quite well — there's a sense of rootedness, people have often lived there for decades, and daily interactions like bumping into each other at the local market keep those connections alive. But in the big cities, especially in the newer high-rise apartment blocks, it's a completely different story — you could live next door to someone for years and never exchange more than a nod in the lift. I think there's also a generational dimension to this: older people tend to invest more in neighbourly relationships, whereas younger people, particularly those who are renting and moving frequently, often don't see the point. Having said that, I did notice a shift during the pandemic — a lot of neighbourhood WeChat groups sprang up, and people who'd never spoken before suddenly started sharing resources. Whether that's lasted, though, is another question entirely.
---
Sample (Band 8.5+)
I'd say it's enormously important — probably more than a lot of people realise until they find themselves in a situation where they don't have good neighbours. The obvious reason is quality of life: if you have neighbours who are noisy, intrusive, or just hostile, it can make your home, which should be your sanctuary, feel stressful and unwelcoming. But I think there's a deeper point too, which is about belonging — good neighbours give you a sense that you're part of a community, not just living in a box. And then there's the purely practical side: in an emergency, the first person who can help you isn't the police or an ambulance — it's the person next door. So when you weigh all of that up, I genuinely believe that having good neighbours can make or break your experience of a place, regardless of how nice the actual property is.
---
Sample (Band 8.5+)
If I had to rank them, I'd put healthcare at the very top — having a hospital or at least a decent clinic within reasonable reach is genuinely non-negotiable for most people, especially families with young children or elderly relatives. After that, I'd say public transport is the real backbone of urban life — without reliable buses or a metro system, a city just grinds to a halt, and people's daily commutes become unbearable. Schools are obviously critical for families, and I'd also argue that green spaces — parks, riverside paths, that kind of thing — are far more important than they're often given credit for; they're not just "nice to have," they're essential for mental health in an otherwise concrete environment. Supermarkets and basic retail are also part of the picture, though I'd put them slightly lower down simply because delivery services have made access less location-dependent than it used to be. So, to summarise, I think the core four are healthcare, transport, education, and green space — everything else is supplementary.
---
Sample (Band 8.5+)
The differences are quite fundamental, I think, and they go well beyond just the range of products. In a small local shop, you tend to get a much more personal experience — the shopkeeper might know your name, remember what you usually buy, even ask after your family. There's a human connection there that you simply don't get in a vast department store or shopping mall, where the interaction is purely transactional. On the flip side, large city centre shops obviously win on variety and often on price — they have the economies of scale to offer discounts that a small independent simply can't match. Another difference I'd point out is the pace: shopping locally is usually quick, functional, something you do on the way home, whereas going into the city centre tends to be more of an event, a planned outing. I suppose there's also a broader question about what kind of city we want to live in — if everyone defaults to the big chains, we lose the character and diversity that small independent shops bring to a neighbourhood.
---
Sample (Band 8.5+)
My instinct is to say yes — in an ideal world, the nearest school should be the default choice. It makes so much sense from a practical standpoint: shorter commutes, more time for sleep and play, friendships that are rooted in the local community rather than scattered across the city. But I think "always" is the problematic word in that question. The reality is that school quality varies enormously, and if the nearest school is genuinely underperforming, I can completely understand why parents would want to look further afield. There are also cases where a child has specific needs — maybe a talent for music or a learning difficulty — that a local school simply isn't equipped to support. So my position would be: the nearest school should be the strong presumption, but there need to be reasonable exceptions, and ideally, the system should be good enough that those exceptions are genuinely rare rather than the norm.
---
Discourse markers as thinking-time devices. Across all six answers, fillers and discourse markers serve dual purposes: they buy the speaker a fraction of a second to think while simultaneously signalling the structure of the argument. Examples:
Progressive specification (moving from general to specific). Each answer follows a funnel structure: open with a general position, then narrow into specific examples or counterpoints. Q3 is a textbook case: "enormously important" (general claim) → "quality of life" (first reason) → "sense of belonging" (deeper reason) → "practical side — emergency" (concrete example). This prevents the answer from feeling like a list and instead creates a coherent, developing argument.
Pivot-and-qualify pattern. Several answers use concessive structures to demonstrate nuanced thinking without losing coherence:
Each pivot is clearly signalled, so the listener never gets lost in the speaker's shift of position. This is the Band 8+ hallmark: holding two ideas in tension without the response feeling contradictory.
Topic-chaining across questions. The examiner will often ask follow-up questions that build on the candidate's previous answer. The samples are designed to leave natural hooks — e.g., Q3 ends with "make or break your experience of a place," which the examiner could pick up with "Can you give me an example of that?" The candidate should be ready to extend any answer by one or two sentences if prompted.
---
Register-flexible vocabulary. The responses shift register fluidly depending on the point being made:
| Register | Example | Context |
|---|---|---|
| Informal / conversational | "that sort of thing," "you know," "I suppose" | Opening gambits, hedging |
| Semi-formal / analytical | "mutual vigilance," "non-negotiable," "economies of scale" | Core argument points |
| Abstract / reflective | "sense of belonging," "the character and diversity that small independent shops bring" | Closing reflections |
This register flexibility — knowing when to sound casual and when to sound analytical — is a key Band 8+ discriminator. The speaker never sounds stiff, but never sounds underprepared either.
Collocation precision:
Paraphrase range. The speaker avoids lexical repetition across answers:
Avoided pitfalls. None of the answers use the Band 6 crutch phrases: "very good," "very important," "more and more." When intensification is needed, the samples use lexical alternatives ("enormously," "genuinely," "fundamentally") or structural alternatives (cleft sentences, rhetorical questions).
---
1. Cleft sentences for focus:
2. Conditional structures with nuance:
3. Comparative structures with embedding:
4. Parenthetical hedging:
5. Ellipsis and fragment completion:
6. Gerund as subject with passive complement:
Error profile. The samples are virtually error-free. Any minor informalities (e.g., "you could live next door to someone for years and never exchange more than a nod") are features of natural spoken grammar rather than mistakes — they mirror the way fluent speakers actually construct sentences in real time.
---
Since these are text-based samples, what follows describes the intended pronunciation features that would characterise a Band 8+ delivery in the IELTS Speaking test.
Connected speech (linking and reduction):
Key phrases that should be delivered as single phonological units:
| Written form | Spoken realisation | Feature |
|---|---|---|
| "sort of a" (Q1) | /sɔːt‿əv‿ə/ | Linking /r/ + schwa reduction |
| "there's a" (Q1) | /ðəz‿ə/ | Contraction + linking |
| "lot of people" (Q2) | /lɒt‿ə ˈpiːpəl/ | Flap /t/ + schwa |
| "don't see the point" (Q2) | /dəʊn siː ðə pɔɪnt/ | /t/ elision in "don't" |
| "kind of city" (Q5) | /kaɪnd‿ə ˈsɪti/ | Linking + schwa |
| "had to rank them" (Q4) | /hæt‿ə ræŋk ðəm/ | /d/ → /t/ assimilation + linking /r/ |
Sentence stress and nuclear tone:
Chunking (sense-group pauses):
The natural pause points in each answer (marked with |) should align with clause boundaries:
Q4: "If I had to rank them, | I'd put healthcare at the very top — | having a hospital or at least a decent clinic within reasonable reach | is genuinely non-negotiable for most people, | especially families with young children or elderly relatives."
The key for Chinese L1 speakers is to resist the impulse to pause between every word. Each sense group should be produced as a single breath unit — practise reading the group aloud in one go, then add the next group.
Chinese L1-specific pronunciation notes:
These are the most common transfer errors from Mandarin Chinese that affect IELTS Speaking pronunciation scores, organised by priority:
| # | Error Pattern | Examples in These Answers | Correction Strategy |
|---|---|---|---|
| 1 | Final consonant dropping | "neighbours" → /ˈneɪbə/ (drops /z/); "range" → /reɪn/ (drops /dʒ/); "help" → /hel/ (drops /p/); "shift" → /ʃɪf/ (drops /t/) | Mandarin has no word-final consonant clusters. Over-articulate the final consonant slightly — hold it for a beat longer than feels natural. For plural /z/ ("neighbours"), make sure the vocal cords are vibrating. |
| 2 | /θ/ and /ð/ → /s/ or /z/ or /f/ | "thing" (Q1), "that" (Q1, heavily repeated), "further" (Q1), "there's" (Q2), "the" (ubiquitous), "whether" (Q2) | /θ/ (voiceless "th"): tongue tip between teeth, blow air. /ð/ (voiced "th"): same position but vibrate vocal cords. The word "that" appears ~16 times across all answers — if every instance is mispronounced as /zæt/ or /dæt/, it creates a persistent L1 accent marker. |
| 3 | Dark /l/ vocalised as /o/ or /ʊ/ | "people" (Q2, Q4), "local" (Q2, Q5), "school" (Q6), "critical" (Q4), "small" (Q5) | Mandarin /l/ is always light (tongue tip to alveolar ridge). English dark /l/ (syllable-final) requires the back of the tongue to rise toward the velum. Practise: say "peop-" and hold the tongue in the velar position — it should sound like a swallowed vowel, not a clear /o/. |
| 4 | Voiced/voiceless final distinction lost | "have" vs. "half"; "live" vs. "life"; "broad" vs. "brought"; "neighbourhood" vs. "neighbourhoot" | Mandarin has no voiced obstruents in final position. "Have" /hæv/ (vocal cords vibrate) must contrast with "half" /hɑːf/ (no vibration). The same goes for /d/ vs. /t/, /z/ vs. /s/, /b/ vs. /p/. Hold the final consonant and check for vocal cord vibration. |
| 5 | Vowel length before voiced/voiceless consonants | "need" /niːd/ (longer vowel) vs. "neat" /niːt/ (shorter); "wide" /waɪd/ (longer diphthong) vs. "white" /waɪt/ (shorter) | English vowels are measurably longer before voiced consonants. In "need to" (Q6), the vowel in "need" should be noticeably longer than the same vowel in "neat." This is a subtle feature but native listeners use it to disambiguate. |
| 6 | Stress-timed rhythm disrupted | Chinese is syllable-timed; English is stress-timed. This means unstressed syllables must be compressed. "combination of practical help" — the unstressed syllables "bination of" and "tical" must be said faster and at lower volume than the stressed syllables. | Clap on the stressed syllables only while speaking: combiNAtion of PRACtical HELP. The unstressed syllables rush to fill the gaps between claps. |
| 7 | Intonation for question-like uncertainty | "Whether that's lasted, though, is another question enTIREly ↗" (Q2) — the rise on "entirely" signals open-ended scepticism. Mandarin uses particles (吗, 呢, 吧) for this; English uses pitch movement. | Practise the rising tone on open-ended statements. If delivered with a flat or falling tone, "is another question entirely" sounds like a closed factual claim rather than an invitation to reflect. |
| 8 | /v/ → /w/ substitution | "very" (Q4), "have" (throughout), "vigilance" (Q1), "services" (Q4) | /v/ = top teeth on bottom lip + vibration. /w/ = rounded lips, no contact. The word "very" cannot become "wery" — this is one of the most immediately noticeable L1 Chinese markers. |
| 9 | Consonant cluster simplification | "strong" → /sɒŋ/; "sprang up" → /spæŋ ʌp/; "children" → /tʃɪldən/; "friends" → /fren/ | Inserting a minimal vowel between consonants is better than dropping one: "sətrong" is more acceptable than "song." But the goal is smooth coarticulation — the /t/ and /r/ in "strong" should overlap, not be separated. |
| 10 | Schwa /ə/ avoidance | "about" → /eɪˈbaʊt/ (full vowel); "a question" → /eɪ/ not /ə/; "facilities" → full /iz/ not /əz/ in rapid speech | Mandarin doesn't have a reduced central vowel. Practise: relax the jaw completely and let the vowel collapse to the centre of the mouth. Every unstressed "a" and many unstressed "e/i/o/u" vowels should reduce to /ə/ in connected speech. |
Priority order for practice: Fix items 1, 2, and 8 first — final consonants, "th" sounds, and /v/ vs. /w/ — because these are the most frequent across all answers and the most immediately noticeable to an IELTS examiner. Items 4, 5, and 6 are more advanced but make the difference between Band 7 and Band 8 pronunciation.
---