How IELTS Writing Is Scored: The Four Criteria Explained

If you understand how IELTS Writing is scored, you stop guessing what examiners want and start practising the things that actually move your band. The short version: a trained examiner marks each of your two tasks against four equally weighted criteria, awards a whole band from 0 to 9 for each criterion, and your Writing band is calculated from those numbers — with Task 2 counting twice as much as Task 1.

The rest of this guide walks through each part of that sentence, because every part of it has practical consequences for how you should prepare.

Who scores IELTS Writing?

Trained, certificated IELTS examiners score every Writing response — including the computer-delivered test. The real exam is never marked by software: a human examiner reads your two tasks and assesses each one against the official band descriptors, criterion by criterion.

Examiners do not judge your writing by general impression. They are trained and regularly monitored to apply the same published standard: the IELTS Writing band descriptors, which define what performance looks like at every band, for every criterion. That is good news for you — the standard is public, stable, and learnable.

What are the four IELTS Writing criteria?

Every IELTS Writing task is marked on four criteria, each worth 25% of the task score:

1. Task Achievement (Task 1) / Task Response (Task 2)
2. Coherence and Cohesion
3. Lexical Resource
4. Grammatical Range and Accuracy

No criterion outweighs another, which has a direct practical consequence: a stunning vocabulary cannot rescue an essay that never answers the question, and perfect grammar cannot rescue a report with no overview.

Task Achievement / Task Response

This criterion asks one thing: did you do the task you were given?

In Academic Task 1 (Task Achievement), that means reporting the key features of the chart, table, or diagram accurately, comparing them where relevant, and — crucially — giving a clear overview of the main trends. In Task 2 (Task Response), it means answering all parts of the question with a clear position that you develop and support throughout.

Common ways students lose marks here: writing about the general topic instead of the specific question asked; giving an opinion without supporting it; omitting the Task 1 overview; and writing under the minimum length — at least 150 words for Task 1 and 250 for Task 2.

Coherence and Cohesion

This criterion measures organisation and flow: logical paragraphing with one central idea per paragraph, clear progression from one idea to the next, and cohesive devices (linkers, referencing, substitution) used naturally.

The classic mistake is treating linking words as decoration. Bolting "Firstly", "Moreover", and "In conclusion" onto every sentence does not demonstrate cohesion — the descriptors explicitly reward cohesion that attracts no attention to itself, and penalise mechanical or faulty use. Linkers should do real logical work or be left out.

Lexical Resource

This criterion measures the range and precision of your vocabulary: topic-appropriate word choice, natural collocation, accurate word formation, and spelling.

Precision beats ambition. A memorised "band 9 word" used slightly wrongly costs you more than a simpler word used exactly right, because the descriptors reward flexible and precise use, not rare words. Repetition, imprecise word choice, and an informal register in Task 2 are the usual sources of lost marks.

Grammatical Range and Accuracy

This criterion measures two things in tension: the variety of structures you use and the density of errors in them. You need both — a wide range full of mistakes scores no better than a narrow range done perfectly.

This is why forcing long, complex sentences usually backfires. A mix of accurate simple sentences and well-controlled complex ones demonstrates more grammatical control than sprawling constructions that collapse halfway through.

How is the band score calculated?

Each criterion receives a whole band from 0 to 9 — there are no half bands at the criterion level. The four criterion bands for a task are equally weighted and averaged, and the average is rounded to the nearest half band. Your final Writing band then combines the two task bands, with Task 2 contributing twice as much as Task 1.

Pairing it with a Task 1 report scored 6.0, the Writing band is (6.0 + 6.5 + 6.5) ÷ 3 = 6.33, which rounds to 6.5.

Rounding follows the convention IELTS publishes for band averaging, and it works in your favour: an average ending in .25 rounds up to the next half band, and one ending in .75 rounds up to the next whole band.

Two practical takeaways follow from the maths. First, because Task 2 counts double, an hour of Task 2 practice buys you more band than an hour of Task 1 practice. Second, because criteria are whole bands, moving your weakest criterion up one band moves the whole average — your fastest route to a higher score is usually your worst criterion, not your best one.

What are the IELTS band descriptors?

The band descriptors are the official marking standard: a public grid, published by IELTS, that describes what performance looks like at each band from 0 to 9 for each of the four criteria — one grid for Task 1 and one for Task 2.

They are the single most under-used free resource in IELTS preparation. If your target is band 7, read the band 6 and band 7 rows side by side for each criterion: the difference between them is a precise description of what you need to change. For example, at band 6 cohesion may be "faulty or mechanical"; at band 7 it "attracts no attention". That is not a vague aspiration — it is an instruction.

Why is my score lower than I expected?

Almost always because of one of these five patterns, each mapping to one criterion:

1. You answered the topic, not the question. The essay is on-topic but never addresses the specific instruction (Task Response).
2. Your linkers are decorative. Every sentence starts with a connector, but the ideas underneath do not progress logically (Coherence and Cohesion).
3. Ambitious vocabulary, imprecise use. Memorised academic phrases deployed where they do not quite fit (Lexical Resource).
4. Complexity without control. Long sentences with an error in every clause (Grammatical Range and Accuracy).
5. Mechanical penalties. Under-length answers, or memorised chunks the examiner is trained to discount.

If you cannot tell which of these applies to your writing, that diagnosis — not more essay-writing — is the thing to fix first. Writing ten more essays with the same undiagnosed weakness practises the weakness.

How our simulator applies the same criteria

We built IELTS Writing Simulator around one rule: the score must come from the official criteria, not from a model's opinion.

When you submit an essay, the language model never assigns your band. Instead, it audits your writing against a structured checklist derived from the official band descriptors — dozens of specific yes/no checks per criterion — and deterministic code computes the band from those results, using the same equal weighting, averaging, and rounding described above. The model cannot be talked into a higher score, because the model does not score.

The whole system is calibrated against essays scored by official examiners, and it is designed never to inflate your result — the feedback you get is the honest signal you need, criterion by criterion.