The official IELTS writing marking scheme is published as a set of band descriptors — short descriptions of what writing looks like at Band 5, 6, 7, 8, and 9 across four criteria. The descriptors are accurate but incomplete. They tell you the criteria but not how examiners apply them in practice, nor what the 0.5-band increments mean, nor which criterion-by-criterion patterns most often separate 7.0 from 7.5. This guide fills those gaps.

The 4 Criteria (and the Weights)

Every IELTS writing test — both Task 1 and Task 2 — is scored against the same four criteria, each worth 25% of the final band:

Task Response (TR) for Task 2, or Task Achievement (TA) for Task 1 — whether you have answered the question and developed your response with relevant, extended ideas.
Coherence and Cohesion (CC) — whether the organisation of your ideas, paragraph structure, and use of linking language helps the reader follow your argument.
Lexical Resource (LR) — the range, accuracy, and appropriateness of your vocabulary, including collocations and paraphrasing.
Grammatical Range and Accuracy (GRA) — the variety of sentence structures you use, and how free your writing is from grammatical errors.

Each criterion is scored on the 0-9 band scale, in 0.5 increments. The final writing band is the average of the four criterion scores, rounded to the nearest 0.5. So if you score 7.0 + 7.0 + 7.5 + 7.0, the average is 7.125, which rounds to 7.0. If you score 7.0 + 7.5 + 7.5 + 7.0, the average is 7.25, which rounds to 7.5. The 0.25-band increments in your criterion scores matter.

The Public Band Descriptors, Plainly Translated

The official descriptors are dense. Here is what they actually mean, criterion by criterion, at the bands where most candidates cluster (6.0, 7.0, 8.0).

Task Response / Task Achievement

Band 6.0: addresses the task, presents relevant main ideas but some may be inadequately developed or may be repetitive. Conclusions may be unclear or repetitive.
Band 7.0: addresses all parts of the task, presents a clear position throughout, presents, extends and supports main ideas but may over-generalise or lack focus in places. Presents a clear conclusion.
Band 8.0: sufficiently addresses all parts of the task, presents a well-developed response with relevant, extended and supported ideas. The conclusion is clear and appropriate.

The single biggest difference between 6.0 and 7.0 here isextension: at 6.0 you mention an idea; at 7.0 you develop it. The difference between 7.0 and 8.0 isfocus: at 7.0 you may have one body paragraph that is less focused; at 8.0 all ideas are clearly developed and relevant. Task Response is the most common "ceiling" criterion for 6.5 candidates, because idea development is the hardest thing to fake under timed conditions.

Coherence and Cohesion

Band 6.0: arranges information coherently with clear overall progression, uses cohesive devices effectively but cohesion within/between sentences may be faulty or mechanical.
Band 7.0: logically organises information with clear progression, presents a clear central topic within each paragraph, uses a range of cohesive devices appropriately though with some under/over-use.
Band 8.0: sequences information and ideas logically, manages all aspects of cohesion well, uses paragraphing sufficiently and appropriately.

Notice what is missing from these descriptors: "uses linking words". The examiners do not score connector variety directly. They score whether the reader can follow your argument. Mechanical connectors ("furthermore, in addition, moreover") actually count against you at higher bands because they signal "I have learned that Band 7 essays use connectors" rather than "I have a clear argument". At Band 8.0, you use logical reference ("this approach", "such a policy") more than connectors.

Lexical Resource

Band 6.0: uses an adequate range of vocabulary for the task, attempts to use less common vocabulary but with some inaccuracy, makes occasional errors in word choice and collocation.
Band 7.0: uses a sufficient range of vocabulary to allow some flexibility and precision, uses less common lexical items with some awareness of style and collocation, may produce occasional errors in word choice and collocation.
Band 8.0: uses a wide range of vocabulary fluently and flexibly to convey precise meanings, skilfully uses uncommon lexical items with occasional inaccuracies in word choice and collocation, produces rare errors in spelling and/or word formation.

The 6.0 → 7.0 jump is about collocation: using "underlying problem" instead of "main problem", "long-term implications" instead of "far-away results". The 7.0 → 8.0 jump is about paraphrase quality: at 7.0 you may still use some prompt language verbatim; at 8.0 every prompt word has been rephrased. The mistake 7.0 candidates make is using rare words in wrong collocations ("make a damage" instead of "cause damage") — this scores lower than simple words used correctly.

Grammatical Range and Accuracy

Band 6.0: uses a mix of simple and complex sentence forms, makes some errors in grammar and punctuation but they rarely reduce communication.
Band 7.0: uses a variety of complex structures with good control of punctuation and grammar, produces frequent error-free sentences, has good control of grammar and punctuation but may make a few errors.
Band 8.0: uses a wide range of structures, the majority of sentences are error-free, makes only very occasional errors or inappropriacies.

The 6.0 → 7.0 jump is about structural variety: at 6.0 you use mostly simple and compound sentences; at 7.0 you use a mix of complex sentences (conditionals, passive voice, participle clauses, relative clauses, cleft sentences). The 7.0 → 8.0 jump is about error rate: at 7.0 you have "a few errors"; at 8.0 you have "very occasional errors". Most candidates at 6.5-7.0 already have enough grammar to reach 7.0; the difference is structural variety, not grammar rules.

How Examiners Actually Apply the Scheme

The public materials describe what the criteria measure. They do not describe how examiners apply them in practice. Three things that experienced IELTS trainers know but the public materials do not emphasise:

Examiners score in one pass. A trained examiner reads your essay once and assigns a band to each criterion based on overall impression. They do not run a checklist. This means the 0.5-band increment comes from the examiner's calibrated sense of "this is mostly 7, sometimes 8" — not from a precise count of features. The implication: trying to game the scheme with feature-counting (e.g. "I need 5 complex sentences") does not work, because the examiner's impression is holistic.
Task Response has a soft floor. If you score below 6.0 in Task Response, your final band is capped at 6.0 regardless of how well you do in the other criteria. The reason: a poorly-answered question makes the rest of the essay irrelevant. This is why a 9.0 in grammar with a 5.0 in TR still scores 6.5-7.0 overall, not 8.0+.
Word count is not scored directly, but a short essay hurts Coherence. The official materials say "no penalty for being under or over the recommended word count". In practice, an essay under 240 words (Task 2) or under 150 words (Task 1) usually cannot demonstrate the paragraph development Coherence measures, so it gets penalised indirectly. A 7.0 candidate writing 230 words is still a 7.0; a 7.0 candidate writing 180 words is usually a 6.5 because the paragraphs are under-developed.

The Patterns That Most Often Cap a Candidate Below 7.0

These are the patterns the public materials do not warn about, but which consistently cap candidates at 6.5 even when their grammar and vocabulary are 7.0+:

The over-generalised example. "Many people believe that..." or "It is well known that..." — these are not examples, they are vague appeals. The 6.5 candidate uses them constantly. The 7.0+ candidate uses a specific, real or plausible-specific, observation: "The shift to remote work between 2020 and 2022 in three major US cities reduced peak-hour transit use by 28%."
The underdeveloped second body paragraph. Body 1 has 6 sentences; Body 2 has 3. The examiner reads Body 2 last and forms the final impression. A short Body 2 reads as "ran out of time" or "did not have a second idea". The 7.0+ candidate plans two reasons of equal depth.
The mechanical conclusion. "In conclusion, I believe that..." followed by a restatement of the introduction. This is the 6.0-6.5 conclusion. The 7.0+ conclusion adds an implication or a forward-looking claim: "If this pattern continues, the cost of inaction will exceed the cost of action within the next decade."
The "studies show" pattern. "Studies show that..." or "Research has found that..." without a specific source. The examiner reads this as either hallucinated evidence or vague bluffing. The 7.0+ candidate either names a likely-real source ("according to OECD data from 2019") or makes the example clearly hypothetical ("Consider a city that...").

How to Use the Marking Scheme in Practice

The most useful application of the marking scheme isnot reading the public descriptors (which most candidates already do). It is scoring your own essays criterion by criterion, then comparing to the official description. Most candidates at 6.0-7.0 score their essays holistically ("this is about 6.5") and miss the criterion where the gap actually is.

A better approach: after writing a Task 2 essay, score each criterion on the 0-9 scale separately. The criterion with the lowest score is the next one to fix. Most candidates spend 80% of their revision time on grammar (because grammar errors are easy to see) and 20% on Task Response (because "my ideas are not developed" is hard to act on). Inverting that ratio is the single biggest lever for the 6.5 → 7.0 jump.

Frequently Asked Questions

Is the marking scheme different for Task 1 and Task 2?

The four criteria are the same, but Task 1 uses "Task Achievement" (TA) instead of "Task Response" (TR) and the emphasis is on overview quality, key feature selection, and accurate data reporting rather than position clarity and argument development. Task 2 carries 2/3 of the writing band, Task 1 carries 1/3.

Are there different marking schemes for Academic and General Training?

The criteria are the same, but General Training Task 1 is a letter-writing task rather than a chart description. The same 0-9 band scale applies; the descriptors are the same; the band score contributes to your overall writing band in the same way.

Can I get a 0.5-band bonus for particularly good work?

No. Bands are awarded in 0.5 increments, but the average of the four criteria is rounded to the nearest 0.5. So a 7.25 average rounds up to 7.5; a 7.125 average rounds down to 7.0. You cannot get a "bonus" beyond what the four criteria support.

How do I know which criterion to focus on for the next 0.5 band?

Score your last 3 essays criterion by criterion. The lowest criterion (averaged across the 3 essays) is your bottleneck. The bottleneck is what to fix next, regardless of which criterion you enjoy working on. Most 6.5 candidates have Task Response as the bottleneck, not grammar.

Are the public band descriptors the same as what examiners use?

The public descriptors are simplified versions of the internal examiner training materials. The internal materials are more granular and include example essays at each band, which the public materials do not. Candidates who have access to the internal materials (through a teacher) often have a small advantage because they can calibrate to actual band-7 essays rather than to the description of a band-7 essay.

Check My Essay Free

Free scan first. No credit card required.

IELTS Writing Marking Scheme Explained: How Examiners Score Your Essay