The band descriptors are the tables examiners use to mark your Writing. IELTS publishes them as a free PDF: one table per task, bands 0 to 9, four criteria. They are dense, so this guide puts bands 6, 7 and 8 into plain English and checks a full essay against them.
Every page of the PDF starts with one rule: "A script must fully fit the positive features of the descriptor at a particular level." You get band 7 on a criterion only when all the band 7 features are there. One strong paragraph doesn't earn it; one weak habit can hold it back.
The 4 criteria, in plain words
Every Writing task gets four scores, and each one weighs the same. The task score is their average: 8, 6, 7 and 7 give 7.0. So half a band gained on your weakest criterion lifts the average just as much as half a band gained on any other.
| Criterion | What the examiner checks |
|---|---|
| Task Achievement (Task 1) / Task Response (Task 2) | Did you do what the task asked, fully and with enough detail? In Task 2: a clear position and ideas that are extended and supported. |
| Coherence and Cohesion | Can the reader follow you? Paragraphs, logical order, linking words and reference words like this and such. |
| Lexical Resource | Vocabulary: range, precise word choice, collocations (words that naturally go together), spelling and word forms. |
| Grammatical Range and Accuracy | A range of sentence types, how many sentences are error-free, and punctuation. |
Why Task 2 counts twice as much
Task 2 contributes twice as much as Task 1. IELTS doesn't publish the exact formula or rounding, so picture it as (Task 1 + Task 2 + Task 2) ÷ 3.
- Task 1 at 8, Task 2 at 5: (8 + 5 + 5) ÷ 3 = 6.0.
- Task 1 at 5, Task 2 at 8: (5 + 8 + 8) ÷ 3 = 7.0.
Same scores, a full band apart. Hence the advice of about 20 minutes for Task 1 and 40 for Task 2, and this guide's focus on Task 2. How to keep to those limits on the day is in our guide to Writing time management. The essay types and a 40-minute plan are in our IELTS Writing Task 2 guide.
Where Task 2 essays lose points
In 73% of Task 2 essays scored on GoPrep, Grammatical Range and Accuracy was the lowest criterion or tied for lowest. Of 14 sub-scores the GoPrep AI examiner gives inside the criteria (GoPrep's own breakdown, not official IELTS), grammatical accuracy is the lowest on average, about 5.5.
| Criterion | Share of essays where it is lowest | Weakest part (average) | Strongest part (average) |
|---|---|---|---|
| Grammatical Range and Accuracy | 73% | Grammatical accuracy: 5.5 | Range of structures: 6.0 |
| Lexical Resource | 57% | Accuracy of word choice: 6.0 | Spelling: 6.0 |
| Task Response | 39% | Developing ideas: 6.0 | Format: 6.5 |
| Coherence and Cohesion | 36% | Linking words: 6.0 | Paragraphing: 6.5 |
Column 2: 3,198 Task 2 essays, 28 August to 26 September 2026; ties count for each tied criterion. Columns 3 and 4: 17,850 Task 2 answers, 28 July to 26 September 2026, rounded to the nearest half band.
Average scores for format and paragraphing were higher than for grammatical accuracy. What changes between band 6 and 7? The GoPrep AI examiner picks up to three priority fixes per answer; here is how often each kind was among them.
| Priority fix | Answers at 6.0 to 6.5 | Answers at 7.0 to 7.5 | What to check in your essay |
|---|---|---|---|
| Subject-verb agreement | 44% | 19% | The verb after every subject: schools do, each child has |
| Run-on sentences and fragments | 17% | 7% | Each sentence has a main verb and ends where the idea ends |
| Word form | 19% | 7% | Noun or adjective: economy or economic |
| Word choice and collocation | 53% | 70% | Words that naturally go together: make progress, not do progress |
| Cohesion and reference | 42% | 58% | What this, it and they point to, and whether each linker fits |
2,458 answers scored 6.0 to 6.5 and 1,111 scored 7.0 to 7.5, out of 8,010 Task 2 answers with priority fixes, 28 July to 26 September 2026. These percentages show how often each issue appeared among up to three priority fixes, not how often it occurred overall. Using only the first answer per learner within each score group changed them by no more than 3 percentage points.
At 6.0 to 6.5, basic grammar still takes many of the three slots; at 7.0 to 7.5 the fixes move to word choice and cohesion. First control, then precision. If your essays sit at 6.5, see what changes from 6.5 to 7 in Writing.
Band 6 vs 7 vs 8 on each criterion
The Task 2 descriptors in plain English. The exact wording is in the official PDF.
| Criterion | Band 6 | Band 7 | Band 8 |
|---|---|---|---|
| Task Response | Covers the main parts of the question, some more fully than others. A position is there, but conclusions may be unclear or repetitive. Some ideas aren't developed enough. | Covers the main parts of the question with a clear position. Main ideas are extended and supported, but may over-generalise or lack focus. | Addresses the question appropriately and sufficiently. A clear, well-developed position; ideas are well extended and supported, with occasional lapses at most. |
| Coherence and Cohesion | A clear overall progression, but linking can be faulty or mechanical. Paragraphing isn't always logical. | Logically organised, with clear progression. A range of linking and reference words used flexibly, with some errors or over/under use. | Easy to follow and logically sequenced. Paragraphing is sufficient and appropriate; occasional lapses only. |
| Lexical Resource | Adequate vocabulary, but a restricted range or imprecise word choice. Taking risks brings more range and more errors. | Enough range for some flexibility and precision. Some less common words, and awareness of style and collocation, with some slips. | A wide range used fluently for precise meanings. Skilful use of uncommon and/or idiomatic expressions where appropriate, with occasional slips in word choice or collocation. |
| Grammatical Range and Accuracy | A mix of simple and complex sentences, with limited flexibility. Complex sentences have more errors; errors rarely block meaning. | A variety of complex structures. Error-free sentences are frequent; a few errors may remain. | A wide range of structures. The majority of sentences are error-free; occasional, non-systematic errors. |
Read across the rows and one idea repeats. 6 to 7 is about control: the band 6 writer has ideas and words but can't keep them working. 7 to 8 is about precision: the band 8 writer says exactly what they mean.
The examples below use one discussion essay question. The examiner scores the whole essay, so treat each pair as a hint, not a sample with a guaranteed band.
Some people think children should start learning a foreign language at primary school, while others think it is better to start at secondary school. Discuss both views and give your own opinion.
Task Response: from naming an idea to developing it
At band 6, some main ideas "may be insufficiently developed". At band 7 they are "extended and supported", and at band 8 "well extended", without the band 7 "tendency to over-generalise".
Named, not developed
Learning a foreign language at primary school is good for children because it helps them in the future.
Helps how, and why does the age matter? The reader can't tell.
Extended and supported
Starting at primary school can give children several extra years of practice before university, so they are more likely to use the language with confidence when their studies require it.
Extended, with a limit
The benefits of this head start may be limited, however, if primary schools lack teachers who can provide accurate language models.
The third sentence continues the second. Instead of claiming that an early start always works, it says when the benefit may be smaller: no over-generalising.
Coherence and Cohesion: from stacking linkers to connecting ideas
At band 6, cohesion may be "faulty or mechanical" because of misuse or overuse. At band 7, linking and reference words are used flexibly.
Mechanical
Firstly, children learn fast. Secondly, they have more time. Moreover, it is good for their future. Furthermore, parents like it.
Connected
Starting at primary school gives children more years to practise. This head start matters because practice adds up: a child who starts at seven and keeps learning until seventeen can have ten years of practice.
This head start points back to the previous idea, instead of a linker glued to the front of each sentence. More in linking words for Task 2.
Lexical Resource: precise, not just rare
The band 6 descriptor warns that if the writer is a risk-taker, "there will be a wider range of vocabulary used but higher degrees of inaccuracy or inappropriacy".
Risky
Early language acquisition engenders a plethora of cognitive merits for juveniles.
Juveniles sounds legal where children is natural, and engenders a plethora of merits is not a natural collocation.
Precise
Many parents see an early start as a clear advantage, especially for pronunciation.
Common words, used exactly. Band 7 asks for "some flexibility and precision", and band 8 adds "uncommon and/or idiomatic items" where they fit. Idiomatic means natural fixed phrases like a head start, not colloquial idioms like a piece of cake: don't insert those. Which synonym fits which sentence is covered in Task 2 vocabulary by meaning.
Grammatical Range and Accuracy: count your error-free sentences
A clear line between the bands is how many sentences have no errors. At band 7, error-free sentences are "frequent"; at band 8, "the majority of sentences are error-free". At band 6, complex sentences are less accurate than simple ones.
A complex sentence that breaks
Although starting early has benefits, but many schools does not have teachers which can teach it well.
Three errors: although and but together, schools does, and which for people.
The same idea, under control
Although starting early has benefits, many schools do not have teachers who can teach a language well.
A complex sentence you control helps you. One that collapses counts against you.
A full essay checked against the 4 criteria
A complete answer to the same question, written by the GoPrep team as a practice example. It is not an officially marked script, so we don't give it a band (about 260 words).
Some people believe that children should begin learning a foreign language at primary school, while others argue that it is better to wait until secondary school. Both views have merit, but I believe an early start is more useful, provided that schools can teach the language well.
Supporters of a later start make two reasonable points. First, older children can already read and write well in their own language, so they understand grammar rules more easily. Second, secondary schools usually have specialist language teachers, while a primary school may have to rely on a class teacher with little training. If lessons are poor, young learners may simply memorise mistakes.
On the other hand, starting at primary school gives children a head start: more years to practise. This matters because practice adds up. A child who starts at seven and keeps learning until seventeen can have ten years of practice, so they are more likely to use the language with confidence at university or at work. Younger children also tend to worry less about making mistakes when they speak.
In my view, the extra years of practice outweigh the advantages of waiting, but only when the teaching is good. The benefits of this head start may be limited if primary schools lack teachers who can provide accurate language models. Governments that introduce languages early should therefore train primary teachers first.
In conclusion, waiting until secondary school has real advantages, but an early start, supported by well-trained teachers, gives children the best chance of becoming confident users of a foreign language.
| Criterion | What the essay does |
|---|---|
| Task Response | Each view has its own paragraph. The opinion is stated in the introduction, argued in paragraph 4 and repeated at the end. |
| Coherence and Cohesion | One job per paragraph. Few linkers; This matters and this head start carry ideas between sentences. |
| Lexical Resource | Common words used exactly: specialist language teachers, accurate language models, and one natural fixed phrase, a head start. |
| Grammatical Range and Accuracy | Complex structures under control: If lessons are poor, provided that, clauses with who and that. |
What can hold your score down
Some lines in the PDF are in bold: "negative features that will limit a rating". Fix these first.
Task 1: same four criteria, one different
The difference is Task Achievement: it checks how you report information, not how you argue. In Academic Task 1, band 6 says "a relevant overview is attempted"; band 7 needs "a clear overview", sensibly grouped data and the main trends or differences. Examples for each chart type are in how to write a Task 1 overview. In General Training, the letter covers all 3 bullet points with a clear purpose and a fitting tone (how to write the General Training letter). Task 1 needs at least 150 words in about 20 minutes. More in our guides to the line graph, bar chart, table and Task 1 vocabulary.
How to use the descriptors on your own essay
Check your last Task 2 essay once for each criterion. A self-check shows what to fix, not your band. For scores on all four criteria, see where to get Writing feedback online.
Then pick one problem that repeats and, in your next essay, check for that one first. One habit is easier to control than ten.
Practice: which fix comes first?
A weak body paragraph on the same question. Choose the fix that would help most, then read the answer.
Firstly, children in primary school learn languages very fast. Moreover, they have a lot of time. Furthermore, it is good for there future, because many jobs needs English. So learning early is the best way for all children.
- A. Swap simple words for advanced ones: fast → expeditiously, good → advantageous.
- B. Develop one idea (why more years of practice help) and fix there and jobs needs.
- C. Add more linking words so every sentence connects to the one before.
Answer: B. Three ideas are named and none is developed, the band 6 Task Response problem. A is the risk-taker trap; C makes cohesion more mechanical. Paragraph 3 of the full essay shows fix B in action.
FAQ
What are the IELTS Writing band descriptors?
The official tables examiners use to mark IELTS Writing: bands 0 to 9 for four criteria, one table for Task 1 and one for Task 2, free on ielts.org.
How is the IELTS Writing score calculated?
Each task gets a band for four equally weighted criteria, and the task score is their average. Task 2 counts twice as much as Task 1.
What is the difference between band 6 and band 7 in IELTS Writing?
At band 7, main ideas are extended and supported, linking is flexible and error-free sentences are frequent. At band 6, some ideas are underdeveloped, linking can be mechanical and complex sentences have more errors.
How can I get 7+ in IELTS Writing?
Aim for band 7 features on all four criteria, especially in Task 2: extend and support each main idea, link with reference words rather than a linker on every sentence, and make error-free sentences frequent. You don't need exactly 7 on each criterion: 8, 6, 7 and 7 average 7.0 for a task.
Do IELTS examiners count every mistake?
No. The descriptors set no fixed number of errors. They describe how often sentences are error-free, whether errors block meaning and whether they are systematic.
Is band 7 a good score in IELTS Writing?
On the IELTS scale, band 7 is labelled "Good": operational command of English with occasional inaccuracies. Whether it is enough depends on what your university, employer or visa programme requires.
Is band 8.5 possible in IELTS Writing?
Yes. Scores are reported in whole and half bands up to 9, so 8.5 is possible. It needs criterion scores close to band 9: for example, 9, 9, 8 and 8 average 8.5 for one task. Your Writing band then depends on both tasks.
Where to go next
See how the IELTS band score is calculated and the IELTS Speaking band descriptors. To spread band 7 practice over the time before your exam, tell the quiz your target score, exam date and hardest part of IELTS, and GoPrep will build a personal study plan around your answers.