Free — no sign-up
IELTS band descriptors, explained
What each criterion actually measures, and the two published rules that cap a band whatever else you did well.
Format checked against IELTS on .
They apply to two sections, not four
Reading and Listening are marked out of 40 and converted with a raw-score table — no judgement, no descriptors. The band descriptors are for Writing and Speaking, the two sections a human examiner scores.
Which is why those two are the sections where knowing the criteria changes your score. You cannot argue with a Reading key; you can absolutely write an essay that satisfies a descriptor you had not read.
Writing — four criteria
- Task Achievement (Task 1) / Task Response (Task 2) — did you do the thing asked, completely. Same slot in the descriptors, different question for a different task.
- Coherence and Cohesion — paragraphing and how ideas connect. Not a count of linking words; a paragraph stuffed with moreover and furthermore can score lower than one with none.
- Lexical Resource — range and precision. An ambitious word used wrongly costs more than a plain word used correctly.
- Grammatical Range and Accuracy — variety of structures, and how often they are error-free. Both halves matter: flawless simple sentences cap out, and ambitious broken ones do too.
Speaking — four criteria
- Fluency and Coherence — keeping going, and whether it hangs together. Pausing to think about the idea is treated differently from pausing to hunt for the word.
- Lexical Resource — including paraphrase when the exact word will not come.
- Grammatical Range and Accuracy.
- Pronunciation — intelligibility, not accent. The descriptors ask whether you are understood without strain. No candidate is marked down for sounding like where they are from.
The two caps worth knowing before you sit
No overview in Task 1 caps Task Achievement at band 5. However accurate the rest is. An overview is a sentence or two naming the main trend or biggest contrast, without numbers.
Not answering the question asked caps Task Response in Task 2, regardless of how well the essay is written. Answering half of a two-part question is the common version of this.
Neither is a penalty bolted on. It is what the descriptor at that band says, which is why reading them is worth twenty minutes.
The rounding rule most candidates get wrong
The four section bands are averaged and rounded to the nearest half band — with one twist IELTS publishes and almost nobody quotes:
An average ending in .25 rounds UP to the next half band. An average ending in .75 rounds UP to the next whole band.
So 6.5, 6.5, 6 and 6 average 6.25 and are reported as 6.5, not 6. Work yours out on the IELTS band score calculator, which applies exactly this rule.
Questions, answered
What are the IELTS band descriptors?
They are the public criteria examiners mark Writing and Speaking against. Writing uses Task Achievement or Task Response, Coherence and Cohesion, Lexical Resource, and Grammatical Range and Accuracy. Speaking uses Fluency and Coherence, Lexical Resource, Grammatical Range and Accuracy, and Pronunciation. Each is scored 0 to 9 and the four are averaged for the section band.
How is the overall IELTS band score calculated?
Average the four section bands and round to the nearest half band, with one rule most people miss: an average ending in .25 rounds up to the next half band and an average ending in .75 rounds up to the next whole band. So 6.5, 6.5, 6 and 6 average 6.25 and are reported as 6.5, not 6.
Can one criterion cap my IELTS band?
Yes, in two published cases. A Writing Task 1 response with no overview is capped at band 5 for Task Achievement. A Task 2 response that does not address the question asked is capped under Task Response regardless of how well it is written. Neither cap is a penalty added on top — it is what the descriptor for that band says.
Are Reading and Listening marked on the band descriptors?
No. Reading and Listening are marked out of 40 and converted to a band with a raw-score table, so there is no judgement involved. The band descriptors apply to Writing and Speaking, which are the two sections a human examiner scores.
Why does my score not match my expectations on one criterion?
Because the four criteria measure genuinely different things and most candidates are uneven across them. A very common shape is strong Lexical Resource and Grammatical Range with weak Task Response — writing that reads well and answers a question adjacent to the one asked. An overall band hides that; four separate bands show it.
Free to start — four writing and four speaking evaluations a month, no card.
Practise against them
Writing Task 2 and Task 1 are marked on the four Writing criteria; Speaking practice on the four Speaking ones. Reading and Listening use the raw-score tables instead.