IELTS Speaking Part 3: Thinking Out Loud at Length
Problem: candidates survive Parts 1 and 2, then Part 3’s abstract questions (“Has tourism changed?”, “Why do cities need nature?”) expose thin ideas and present-tense-only grammar. Answer in brief: treat Part 3 as four discursive moves — deepening opinions, comparing past and present, speculating about the future, and evaluating both sides — each with its own language kit. Method, models and 12 tasks below.
Learning outcomes
By the end you will be able to: extend Part 3 answers to 4–6 sentences using the opinion–reason–evidence–balance shape; compare past and present with correct tense contrast; speculate with conditionals and future forms; evaluate issues (tourism, nature, city life) from two sides with concession language; and link Part 3 answers back to your Part 2 topic without repeating it. All guidance follows the source book, Speak Smart: The Complete IELTS Speaking Preparation Toolkit (4 IELTS SP BOOK) (Jobin Thomas), whose speaking introduction, fifty-card Part 2 bank with brainstorming and samples, Part 3 discussion sets and full exam-question sets were verified by full-structure scan with set sampling.
Move 1: deepening opinions with reasons and evidence
Part 3 answers must run longer than Part 1’s 2–4 sentences: aim for four to six, built as opinion–reason–evidence–balance. “Spending time outdoors is essential (opinion). Natural settings drive physical activity and supply vitamin D for bone and immune health (reason, precise). Mentally, green time cuts cortisol and restores attention fatigued by urban overstimulation (second reason, technical lexis). Nature also feeds creativity — many people report their best ideas arriving on walks, not at desks (evidence by observed pattern). In a world of prolonged indoor hours, that makes outdoor time maintenance, not luxury (closing evaluation).”
Evidence comes in four cheap forms when personal experience runs out: observed patterns (“many people…”), plausible mechanisms (“exposure to sunlight provides…”), comparisons (“unlike gym exercise…”), and concessions-as-evidence (“even sceptics concede…”). Rotate them; the same evidence type four answers running sounds rehearsed.
Move 2: past–present comparison
“How has X changed?” questions demand tense contrast done cleanly: past simple/used-to for then, present perfect/present for now, with the turning point named. “Tourism has changed enormously (verdict). Two decades ago it meant package tours booked through agents months ahead (past). Now platforms let travellers assemble trips overnight, which has democratised travel but also concentrated crowds in a handful of photogenic spots (present + perfect + evaluation). The turning point was the smartphone — maps, reviews and bookings in one pocket (mechanism).” The evaluation tail (“democratised… but concentrated”) is what lifts description into discussion — always add what the change means.
Move 3: future speculation
“What will happen?” questions score conditionals and future range: first conditional for likely outcomes (“if cities keep densifying, pocket parks will become…”), second conditional for hypotheticals (“if tourists were capped, revenues would…”), and hedging modals (may, might, is likely to) for calibrated claims. “Overtourism will likely force regulation — permits, caps, seasonal pricing (prediction with hedge). If Venice-style entry fees spread, day-trippers may fall while longer stays rise, which would suit residents (conditional chain). Cities that fail to act, though, will simply export their crowds to the next unprotected town (warning close).” Three future forms in one answer displays the range Part 3 exists to test.
Move 4: two-sided evaluation
Problem/solution and advantage/disadvantage questions need concession architecture: acknowledge the other side in a subordinate clause, then land your verdict in the main clause. “Although tourism strains water, waste and housing (concession, real), its revenues fund the preservation that poverty would abandon — restored quarters exist because visitor fees pay their masons (verdict with evidence). The policy answer is therefore management, not rejection: caps, local ownership and behaviour codes (solution). Tourists can be nudged toward respect through brief orientation, visible codes and pricing that rewards longer, gentler stays (concrete measures).” Subordinate-the-concession, main-clause-the-verdict: the grammar itself performs the evaluation.
Worked models: outdoors and tourism discussions
Labelling: all answers below are newly written for this article, following the book’s set patterns (outdoors importance, travel benefits, tourism change/problems/respect). The book’s questions and sample answers are not reproduced. Suggested models — other valid answers exist.
Model 1 — Outdoors importance, deepened (newly written)
Q: How important is spending time outdoors? “Essential for body and mind alike (opinion). Physically, outdoor settings invite movement — walking, cycling, play — which protects against the chronic diseases of sedentary life (reason). Sunlight’s vitamin D supports bone and immune health, a free supplement most indoor workers underuse (precise mechanism). Mentally, green exposure lowers stress hormones and restores the attention that screens fragment (second reason). Even creativity benefits — the break from overstimulation lets ideas surface (evidence pattern). Given today’s prolonged indoor hours, outdoor time is maintenance, not leisure (evaluation close).” Examiner note: six sentences, four evidence types implied, technical lexis (cortisol-free but precise: chronic, immune, overstimulation); balance embedded in “not leisure”.
Model 2 — Tourism change, compared (newly written)
Q: How has tourism changed in recent years? “Beyond recognition (verdict with voice). It used to mean brochures, agents and fixed itineraries booked months ahead (past). Now travellers build trips overnight on platforms — flights, stays, experiences in one flow — which has made travel spontaneous but also herded crowds into the same few squares (present + perfect + evaluation). The smartphone caused it: navigation, translation and reviews collapsed the old barriers together (mechanism). The next shift is already visible — cities pricing or capping entry, which may redistribute crowds if neighbours coordinate (speculative tail).” Examiner note: tense contrast clean; mechanism named; answer ends looking forward — discussion, not report.
Model 3 — Tourism problems and respect, evaluated (newly written)
Q: What problems can tourism cause, and how can tourists respect local cultures? “The costs are water stress, waste spikes and housing converted to short lets that hollow neighbourhoods (problem list, concrete). Although these are real, abandoning tourism would strand the conservation and craft incomes it funds (concession + verdict). So the answer is managed respect: brief arrival orientations stating local norms, visible codes in stays and sites, and pricing that favours longer stays over hit-and-run day trips (solutions, concrete). Where residents co-own the tourism offer — homestays, guided walks — respect follows naturally, because guests meet hosts, not backdrops (mechanism close).” Examiner note: concession architecture exact; solutions specific enough to picture; close generalises (mechanism) — the abstract move Part 3 rewards.
Model 4 — Linking back to Part 2 without repeating (newly written)
Part 2 was a memorable holiday; Part 3 asks: “Do people travel too much nowadays?” “My Kyoto trip suggests the question is mis-aimed (bridge — references Part 2 without retelling it). The issue isn’t volume but distribution: a dozen dispersed trips teach more and damage less than one honeypot repeated (reframe). Frequent travel broadens those who engage — language attempts, local food, unplanned streets — while conveyor tourism teaches little however often repeated (evaluation with contrast). So too much of the wrong kind, yes; too much of the attentive kind is barely possible (balanced verdict).” Examiner note: bridge-plus-reframe technique — one clause of continuity, then new argument; never re-narrate the long turn.
Rapid-fire model bank: three more topics (newly written)
Technology — “Has social media changed friendship?” “In structure more than feeling (opinion). Before, friendships decayed with distance — letters slowed, then stopped (past). Now feeds preserve weak ties indefinitely, which has widened circles while thinning their centres (present + perfect + evaluation). The change that matters is ambient awareness: we know everyone’s headlines and nobody’s chapters (mechanism with image). If platforms ever priced attention honestly, usage would halve — which tells you the current shape serves advertisers first (conditional verdict).” Note: past–present–mechanism–conditional in five sentences.
Education — “Should university be free?” “Ideally yes, practically with conditions (qualified opinion). Free access widens opportunity — talent is evenly spread, fees are not (reason with aphorism). Although graduates earn premiums that justify contribution, upfront fees deter exactly the students who would gain most (concession + verdict). The workable model is therefore deferred contribution: study free, repay through earnings above a threshold (solution). Countries running variants of this show participation rising without quality collapsing (evidence gesture).” Note: concession architecture plus concrete policy — evaluation at full stretch.
City life — “Will cities keep growing?” “Nearly certain in the developing world, less so where remote work sticks (split prediction). Cities used to monopolise opportunity — jobs, hospitals, universities in one place (past). Now connectivity has unbundled work from workplace, which has slowed megacity growth without stopping urbanisation itself (present + perfect). If transport pricing reflected true costs, mid-size towns would absorb the next wave — planners should build for that scenario now (conditional + recommendation).” Note: split verdicts (“nearly certain… less so”) display the calibration examiners reward.
Practice bank: 12 tasks with answers
Attempt aloud before opening each answer. All items are newly written; the book’s questions and answers are not reproduced.
1 (Deepen): Extend “Parks matter” to four sentences with two reasons. — Show model
Model: “Parks are a city’s lungs and living rooms (opinion with image). They cool heat-island streets and absorb storm water (reason 1, mechanism). They also stage the casual encounters — morning walkers, evening cricket — that knit strangers into neighbours (reason 2, evidence pattern). A city without them isn’t denser, just meaner (evaluation close).” Note: image + mechanism + pattern + verdict is a complete deepen.
2 (Deepen): Give two evidence types for “reading broadens minds”. — Show answer
Answer: mechanism (narrative simulates other lives, building empathy circuits) + observed pattern (readers sustain attention longer in discussion). Note: rotate evidence types across answers.
3 (Compare): Contrast shopping past vs present with a named turning point. — Show model
Model: “Shopping used to mean Saturday markets and known shopkeepers (past). Now one-click delivery has made it solitary and instant — convenient, but the street loses its theatre (present + evaluation). The smartphone checkout caused it (turning point).” Note: verdict, past, present, mechanism, meaning — five beats.
4 (Compare): What tense error ruins past–present answers? — Show answer
Answer: present-tense past (“before people travel less”) or past-tense present (“now it was convenient”). Note: used-to/past for then, present perfect/present for now — drill the pairing.
5 (Speculate): Predict remote work’s effect on cities with two future forms. — Show model
Model: “If hybrid work holds, commuter towns will hollow on weekdays and bloom on weekends (first conditional + future). Were offices to shrink permanently, city centres would convert to housing — which might finally cool rents (second conditional + hedge).” Note: likely + hypothetical in one answer shows full range.
6 (Speculate): Hedge this flat prediction: “Tourism will destroy old towns.” — Show fix
Fix: “Unmanaged tourism is likely to erode old towns — though capped, priced entry may preserve them.” Note: hedge + condition converts alarm into analysis.
7 (Evaluate): Concede-then-verdict on “social media harms friendship”. — Show model
Model: “Although feeds displace face time (concession), they also sustain the weak ties — school friends, distant cousins — that distance once dissolved (verdict). Net effect: friendships wider, some thinner (balanced close).” Note: subordinate the concession, main-clause the verdict.
8 (Evaluate): Propose three concrete measures for respectful tourism. — Show answer
Answer: arrival orientations on local norms; visible codes in stays/sites; pricing favouring longer stays. Note: concrete beats pious — picture-test every proposal.
9 (Bridge): Part 2 was “a person you admire”; Part 3 asks about role models generally. Bridge in one clause. — Show bridge
Bridge: “My admiration for my teacher suggests the wider pattern — …” then new argument about systems vs individuals. Note: continuity without retelling — the examiner heard the story already.
10 (Length): Your Part 3 answers run two sentences. Diagnose. — Show answer
Diagnosis: opinion without moves — add one reason, one evidence, one balance. Note: length follows structure; the four moves are the length.
11 (Lenses): Blank on “Should zoos exist?” — which move rescues you? — Show rescue
Rescue: evaluate — “Although captivity constrains, well-run zoos fund the field conservation that habitat loss outpaces; the verdict depends on the zoo, not the concept.” Note: evaluation converts any dilemma into structure.
12 (Mixed, timed): 2-minute drill: one past–present–future answer on “public transport”. — Show plan
Plan: past — deregulated chaos, used-to; present — app-tracked but crowded (perfect); future — if fares integrate, ridership will jump; verdict — coordination, not vehicles, decides. Note: all three time zones plus a verdict in six sentences.
Common errors and corrections
- Part-1-length Part 3: two-sentence answers to abstract questions. Correction: four moves to 4–6 sentences — structure is length.
- Present-tense-only: no past or future forms displayed. Correction: comparison and speculation moves force the full tense range.
- Description without meaning: reporting change without evaluating it. Correction: every comparison ends with what it means.
- Unhedged alarm: “will destroy” predictions. Correction: hedge + condition — analysis, not alarm.
- Concession-free verdicts: one-sided argument. Correction: subordinate the other side, main-clause your verdict.
- Part 2 retelling: re-narrating the long turn. Correction: bridge in one clause, then new argument.
Independent task and self-check
Task (newly written): take three Part 3 questions from the book’s exam sets and record six-sentence answers, each using a different lead move (deepen / compare / speculate / evaluate). Transcribe and label every sentence’s function.
Self-check: does every answer reach 4–6 sentences? Are past, present and future forms all present across the three? Is one concession correctly subordinated? Is Part 2 referenced at most once? Four yeses = discussion mastery.
Study sequence with the book
- Read the speaking introduction; fix Part 3’s purpose (abstract discussion, 4–5 minutes).
- Study five Part 3 models per topic, labelling each sentence’s move.
- Drill one move weekly (deepen → compare → speculate → evaluate) with this article’s tasks.
- Attempt full exam-question sets, bridging Part 2 to Part 3 without retelling.
- Record monthly; count sentence functions and tense range, not just fluency feel.
- Final loop: random Part 3 questions, six sentences, move-labelled — retire a topic after three clean attempts.
FAQs
How long should Part 3 answers be? Four to six sentences — developed but not monologues. The examiner needs chances to probe; overlong answers steal your own follow-up opportunities.
Can I admit ignorance on abstract questions? Briefly, then argue anyway: “I have never studied this, but the logic seems…” — one honest clause, then the moves. Silence is the only failure.
Should Part 3 contradict my Part 2? No — broaden, don’t reverse. New angles consistent with the story read as thought; reversals read as invention.
Which move matters most? Evaluation — concession architecture is the single fastest band-raiser because it forces complex grammar and balanced lexis simultaneously.
How does this book differ from the other Speaking books? Frameworks covers three-part shapes; Deep Dive drills Part 2 brainstorming; 50-Sets trains full simulation; Topic Bank covers Part 1; Myths dissects long-sample architecture; this toolkit volume contributes Part 3 discussion depth plus full exam-question sets. Complete the sequence in that order.
The matching book
Speak Smart: The Complete IELTS Speaking Preparation Toolkit (4 IELTS SP BOOK) contains the full apparatus: the speaking introduction, fifty Part 2 cue cards with brainstorming and band-9-style samples, Part 3 discussion sets (outdoors, travel, tourism and more), and full exam-question sets for mock practice. If discussion depth helped here, the book is where you drill it across every topic. Product page link to be added when the authenticated staging catalogue mapping is available — no URL is invented here.
Keep practising with Speak Smart: The Complete IELTS Speaking Preparation Toolkit (4 IELTS SP BOOK)
Enjoyed this guide? Speak Smart: The Complete IELTS Speaking Preparation Toolkit (4 IELTS SP BOOK) takes you further, with deeper explanations, more practice tasks and complete model answers. Continue practising with the full book →
Sources and editorial note
Teaching follows the verified source EPUB (introduction ~4,000 words; 50-card Part 2 bank with brainstorming; Part 3 discussion sets; exam-question sets, confirmed by structural scan and set sampling). All answers in this article are newly written in the book’s patterns; no book question, sample or phrase is reproduced, and no third-party material is copied. New answers are suggested models. Test-format summaries are editorial paraphrase; current exam rules should be confirmed against the official IELTS documentation before booking.
Put this into practice
Compare these course outlines and choose the practice that matches your next goal.
