Why GT Task 2 is predictable — and why that doesn't mean memorizing

Task 2 is worth twice Task 1's marks and examiners score the same four criteria every time: Task Response, Coherence & Cohesion, Lexical Resource, Grammar. The topics repeat because the test measures argument, not subject knowledge. Multiple 2026 trackers note the same thing: examiners are actively cracking down on memorized template essays. The play is not a script — it's a reusable position formula you can run on any topic.

The position formula (runs on every question type)

Position → Reason → Example → Widen.

  • Position: answer the question directly in one sentence. "I believe governments should fund public transport first."
  • Reason: one cause-effect chain. "Because congestion costs cities more than new roads save."
  • Example: one specific, real-world anchor — your country, your city, a stat you actually know. Examiners trust specific over grand.
  • Widen: connect it outward one step. "The same logic applies to buses in rural towns."

Four question types, one formula — the identification drill is in the published guide IELTS Academic Writing Task 2: the four question types and the formula for each (works identically for GT).

The 12 topic clusters showing up in 2026

Cross-checked against 2026 topic trackers and a reported February 2026 Canadian GT sitting. Each cluster comes with one position seed — a starting angle, never a script.

1. Education vs experience — employers value degrees or skills? Seed: argue the stage matters (degrees open doors, skills close the salary gap). 2. Online learning vs classroom — "will replace classrooms" appears yearly. Seed: replace ≠ remove; hybrid absorbs the best of both. 3. Technology's double edge — smartphones, social media, screen time. Seed: the tool is neutral; regulation and habit design are the real variable. 4. Work and remote work — work-from-home, work-life balance, uniforms, 4-day weeks. Seed: productivity moved from presence to output — argue from that. 5. Environment and climate responsibility — who's responsible: individuals or governments? Seed: individuals adapt fastest, but only governments scale. 6. Public transport vs private cars — the classic; a reported 2026 prediction is "invest in public transport vs building new roads." Seed: argue the cheapest kilometre, not the nicest vehicle. 7. Government spending priorities — health, education, arts, space. Seed: rank by irreplaceability — what only the state can do. 8. Health and lifestyle — fast food, exercise, ageing populations. Seed: prevention is cheaper than treatment — run the money argument. 9. Immigration and multiculturalism — benefits vs challenges of newcomers. Seed: argue with one city-level example (yours, if you have one). 10. Cultural change — cultures becoming similar, celebrities speaking on public issues. Seed: argue what's gained first — counters the default doom position. 11. Advertising and consumerism — are consumers manipulated? Seed: influence ≠ control; the literacy gap is the story. 12. Crime and punishment — stricter sentences vs rehabilitation. Seed: argue from the repeat-offender lens, not the single-crime lens.

Actual recent Canadian GT prompt for calibration (Feb 2026): "Some people prefer staying in hotels when travelling, while others prefer renting holiday homes for a short period of time. Discuss both views and give your own opinion." Notice: discuss-both + opinion = run the formula twice, then plant your position.

The three mistakes that cap Task 2 at 6.5

1. Answering half the question. "Discuss both views and give your opinion" needs three moves, not one. Underline every verb in the prompt before you write. 2. Examples without a point. One detailed specific example beats three vague ones ("many countries..."). If you can't name it, don't use it. 3. Memorized scaffolding. Phrases like "with the development of society" and "every coin has two sides" are template markers — the current crackdown targets exactly these. Keep linking language natural and varied.

How this connects to practice

Every Task 2 on this site's writing practice page runs the same four criteria as the exam. Write one cluster per day, get the AI band estimate, and track which criterion holds you back — that's your syllabus, not more topics.

Fact-check anchor: 12 topic clusters verified Oct 7 2026 against 2026 GT prediction lists and a Jan–May 2026 Task 2 topic analysis; Task 2 = double Task 1 marks; the four scoring criteria (Task Response, Coherence & Cohesion, Lexical Resource, Grammar); memorized-template crackdown confirmed across multiple 2026 sources. The hotels-vs-holiday-homes Feb 2026 Canada prompt is reported, not independently verified — accommodation/travel is a long-standing recurring theme. Sets rotate; re-verify before quoting.