Every IELTS candidate collects model essays, and almost none are helped by them. The problem is not the essays — most band-9 samples are genuinely excellent. The problem is what you can do with an excellent essay written by someone else: admire it, memorise pieces of it, and reproduce neither. This lesson explains the three ways model essays fail you, and what a sample that actually helps looks like.
Failure 1: the grammar is above your control
A band-9 sample is written in band-9 grammar — nested clauses, precise hedging, idiomatic phrasing that a professional produces without thinking. Under exam pressure, you will not produce it by having read it. Worse, trying to produce it is actively expensive: Grammatical Range and Accuracy rewards error-free sentences, and a band-9 structure attempted at band-6 control generates errors a band-6 structure would not.
Model: Were such policies to be implemented without regard for their distributional consequences, the very inequities they purport to address would likely be entrenched. Exam-day imitation: Were such policies implemented without regard the consequences, the inequities they purpose to address would likely entrenched.
The imitation contains three errors and scores below the plain version the same candidate could have written cleanly: If these policies ignore who pays the cost, they may increase the inequality they are meant to reduce. Reading above your level is fine. Modelling above your level converts other people's fluency into your errors.
Failure 2: examiners are trained to spot recited content
Examiners read hundreds of scripts drawn from the same question pool, so memorised material glows: the fluency changes mid-essay, the phrasing matches scripts they marked last week, the polished paragraph sits beside a hand-made one. The descriptors instruct them to discount memorised language — those sentences simply stop counting as evidence of your ability, and your band rests on the sentences you actually produced.
And recited content ages badly against a live pool. A Task 2 question reported by two test-takers this July asked whether mobile phone use is as antisocial as smoking, and whether phones should be banned in the same places. A memorised "technology is good and bad" essay does not survive contact with that question — it hinges on an analogy that must be examined, not a topic to recite about. The current season's questions are full of these twists, and twists are precisely what memorisation cannot handle.
Failure 3: someone else's examples do not fit your position
A model essay is one writer's position, supported by that writer's examples, in that writer's logic. Borrow the example without the position and the seams show immediately. On the old-buildings question reported repeatedly this June, a memorised example about a famous European cathedral does nothing for a candidate whose actual, defensible view is most old buildings are ordinary, and only genuinely historic ones deserve protection. The borrowed example argues against its own essay. Task Response scores the fit between position, ideas and support — and fit is the one thing you cannot memorise.
What a useful sample looks like
A sample is useful exactly insofar as you can reproduce what it does. That gives two requirements:
Built from your own experiences and ideas. The examples should come from your city, your work, your family — because on exam day, that is the material you will actually have. A sample essay about the traffic in your hometown teaches you to deploy your knowledge; a sample about someone else's is a museum piece. Your idea bank from the last lesson is the raw material: a useful sample shows you those ideas, developed to standard.
Calibrated to your target band — plus half. A sample at target + 0.5 uses grammar you can stretch to control and vocabulary you half-know already. You can study it, close it, and rewrite something honestly comparable. Studying band 9 while targeting band 6.5 is like learning to drive by watching a race: inspiring, unrepeatable.
| Band-9 model, someone else's life | Target + 0.5, your own material | |
|---|---|---|
| Grammar | Above your control → errors when imitated | Stretches your control → reproducible |
| Content | Generic or foreign to you → unmemorable | Yours → already in memory |
| Examiner reaction | Recited passages discounted | Reads as authentic, because it is |
| After studying it | Admiration | An essay you can rewrite tomorrow |
The test of any sample is brutal and simple: close it and rewrite the argument in your own words. If what you produce is a broken imitation, the sample was too far above you. If you cannot remember its ideas at all, the ideas were never yours. If you can reproduce eighty percent — that sample is doing its job.
Your drill (15 minutes)
- Take any model essay you have saved. Read one body paragraph twice, then close it.
- Rewrite that paragraph's argument from memory, in your own English. Compare. Mark every place your version broke — each break is a spot where the model exceeds your current control.
- Now write the same paragraph again, but swap in an example from your own life or city, in sentences you are certain of. Notice it is easier and cleaner: this is what reproducible looks like.
- Pick one corroborated question from this season's list or the recall wall and note which of your experiences could support a position on it. The next lesson prices out how to build that into a full personal bank — both the manual way and the fast way.