Stuck at 6.5 in IELTS Writing? Find the Criterion That Caps You

If your last two or three IELTS Writing results all came back at 6.5, you are standing in the most crowded spot on the band scale. 6.5 is what the arithmetic produces when your four criterion scores hover between 6 and 7: a profile of 7, 7, 6, 6 averages 6.5, and so does 7, 6, 7, 6. It is also the point where general practice stops working, because another full essay polishes all four criteria a little instead of lifting the one holding you down. The way out is not more volume. It is a diagnosis.

Why 6.5 is where progress stalls

Your Writing band is the average of four equally weighted criteria: Task Response, Coherence and Cohesion, Lexical Resource, and Grammatical Range and Accuracy. Below 6, almost any practice helps. At 6.5, the picture changes: two or three criteria are already at 7, and pouring effort into them does nothing for your average. The criterion still sitting at 6 is your entire problem, and until you know which one it is, your practice is a coin toss. Our Band 7 guide explains what the target band looks like; this article is about finding which criterion is keeping you from it.

The complication: your official test report does not show the four sub-scores, only the overall band, so the plateau feels opaque. The fix is to get a criterion-level breakdown on your practice essays and read the four sections below against your own habits. Each criterion caps candidates with its own recognisable symptoms.

Task Response: fluent essays, flat scores

Task Response is the criterion you can fail with good English, which makes it the hardest cap to spot from inside your own writing. The telltale symptoms: your essays feel fluent yet the score never moves. Two-part questions come back with one half answered well and the other half thin. Your conclusion holds a slightly different opinion from your introduction. And your body paragraphs restate the claim in new words instead of taking it anywhere.

The descriptor language shows exactly what the 6 to 7 jump requires. Band 6 accepts a relevant position whose conclusions may become unclear or repetitive, and ideas that are insufficiently developed. Band 7 demands a clear position throughout the response, with main ideas that are presented, extended and supported. Two phrases carry all the weight: 'throughout' means your stance cannot wobble between paragraphs, and 'extended' means each idea must be pushed past its first sentence into a mechanism, an example, or a consequence.

The fix is a reading discipline, not a writing one. Before you write, count the jobs in the prompt: 'discuss both views and give your own opinion' is three jobs, and covering two of them brilliantly still caps you. State your position in one sentence and check every paragraph against it. Then develop fewer ideas further instead of listing more. The Task Response drills train exactly this on exam-style prompts: full-coverage theses, spotting drift, and extending an idea instead of restating it.

Coherence and Cohesion: effort that lowers the score

Coherence and Cohesion is the one criterion where more visible effort can produce a lower mark. The symptoms of this cap: every second sentence opens with 'Moreover' or 'Furthermore', paragraphs drift onto a second topic halfway through, and anyone who reads your work says they lost the thread somewhere, even though every individual sentence is fine. If you have memorised a list of linking phrases and your score has not moved, this is probably your criterion.

The descriptors explain why. Band 6 describes cohesion that may be faulty or mechanical, with referencing that is not always clear. Band 7 asks for logical organisation with clear progression throughout, a range of cohesive devices used appropriately, and a clear central topic within each paragraph. Notice what jumped: not the number of linkers, but the logic they sit on. The examiner is reading for whether they ever have to work out how two sentences relate.

So the fix runs opposite to instinct. Hold each paragraph to one topic, stated in its first sentence, and delete every linker that does not mark a genuine contrast or consequence. Check that every 'this' and 'it' has exactly one possible referent. The coherence drills train the machinery separately: choosing the linker that matches the actual logic, rebuilding paragraphs in the only order that works, and referencing a reader can follow without guessing.

Lexical Resource: word lists that never pay off

The Lexical Resource cap has a distinctive signature: you keep learning vocabulary and it keeps not mattering. Feedback flags word partnerships rather than words, things like 'do a decision' or 'strong traffic'. Your paraphrase of the question shifts its meaning, turning 'children' into 'teenagers'. And when you reach for an impressive word, it lands slightly off, in the wrong register or the wrong form.

The descriptor jump is about accuracy of ambition, not size of vocabulary. Band 6 credits an adequate range and attempts at less common vocabulary 'with some inaccuracy'. Band 7 requires a sufficient range to allow some flexibility and precision, and less common items used with awareness of style and collocation. In other words, Band 6 rewards the attempt; Band 7 only rewards the attempt that comes off. A precise everyday word beats an imprecise rare one every time.

Stop learning single words and start learning partnerships: 'heavy traffic', 'pose a threat', 'take measures', stored and retrieved as chunks. When you paraphrase the question, check the meaning survived, and keep the original word where no natural synonym exists. The vocabulary drills work these three failure points directly: paraphrase without distortion, collocation over decoration, and the right form of the right word.

Grammatical Range and Accuracy: ambition versus control

The grammar cap shows up in one of two opposite ways. Either your long sentences collapse halfway, so your accuracy falls exactly when your ambition rises, or you have retreated into short, safe sentences that are all correct and all simple. Both cap the criterion, because it scores two things at once. A third symptom cuts across both: the same small errors recurring in every essay, usually articles, subject-verb agreement, and the comma splice.

Band 6 uses a mix of simple and complex forms and makes some errors that rarely reduce communication. Band 7 uses a variety of complex structures and produces frequent error-free sentences, with good control of grammar and punctuation. The operative word is 'frequent': the examiner is judging a ratio, and a Band 7 script gets most of its sentences completely clean while still showing range. Being understood despite errors is exactly the Band 6 description.

The fix has two halves. For accuracy, keep an error log and hunt one error type per week; our guide to the six most common grammar mistakes covers the usual suspects. For range, attempt the complex structure you can finish cleanly, not the longest one you can imagine, and place deliberately simple, correct sentences between your complex ones. The grammar drills isolate both halves: spotting errors, building complex sentences that stay correct, and killing the comma splice.

Deliberate practice: the loop that breaks the plateau

Once you know your criterion, the practice loop is short and repetitive on purpose. Write one essay under timed conditions. Grade it and look only at your target criterion's score and feedback. Drill that weakness in isolation for a few sessions, then write again and check whether the number moved. This feels slower than an essay a day, but it is the difference between practising and rehearsing your mistakes.

Difficulty matters too. If you practise on questions far below the level you are aiming at, clean scores tell you nothing; if the questions are wildly hard, everything breaks at once and you cannot see your real weakness. Work at the edge of your ability: the Band 7 question set is pitched for exactly this plateau, so the essays you write there expose the criterion that fails first under realistic pressure.

The plateau at 6.5 is not a talent ceiling. It is an information problem: four numbers averaged into one, with the lowest one hidden. Find it, drill it, and the same effort that has been going nowhere starts moving your band.