IELTS band score calculation: how raw marks become bands

IELTS reports five numbers: a band from 0 to 9 for each of the four modules, and an overall band averaging them. Two entirely different scoring machines produce those numbers. Listening and Reading are counted — 40 questions, one mark each, converted to a band. Writing and Speaking are judged — examiners rate your performance against published criteria. Understanding both machines, and the rounding rule that joins them, turns a mysterious result into something you can plan around.

Listening and Reading: raw marks to bands

Every correct answer is worth one mark with no penalty for wrong answers, giving a raw score out of 40. The test partners publish indicative conversions, with small variations between test versions to keep difficulty fair:

• Listening: about 16/40 for Band 5, 23 for Band 6, 30 for Band 7, 35 for Band 8.
• Academic Reading: about 15/40 for Band 5, 23 for Band 6, 30 for Band 7, 35 for Band 8.
• General Training Reading: about 23/40 for Band 5, 30 for Band 6, 34 for Band 7, 38 for Band 8 — higher thresholds because the texts are less dense.
• Half bands sit between these marks: roughly 26–29 in Listening is Band 6.5, for example.

The planning insight is the marginal question: around the Band 6–7 range, each half band costs roughly three to four extra correct answers. If you score 27 in Reading, you are not 'a band away' from 7.0 — you are three questions away from 6.5 and about seven from 7.0, and your error log will usually show one question type that alone accounts for the gap.

Why practice scores wobble — and when to trust them

Raw scores on practice tests move for reasons that have nothing to do with your ability: passage difficulty varies, some Listening accents suit you better than others, and a single misread instruction can swallow a whole question set. This is why the official conversions shift slightly between versions, and why one practice result — good or bad — proves little. Trust the median of your last three timed attempts rather than the latest one, and treat any single-test jump of more than half a band, in either direction, as noise until repeated. The trend line across weeks is the only practice number worth planning around.

Writing and Speaking: criteria, not counts

Writing is assessed on four equally weighted criteria: Task Achievement (Task 1) or Task Response (Task 2), Coherence and Cohesion, Lexical Resource, and Grammatical Range and Accuracy — and Task 2 carries twice the weight of Task 1 in the module band. Speaking uses four criteria of its own: Fluency and Coherence, Lexical Resource, Grammatical Range and Accuracy, and Pronunciation. Because criteria are averaged, a single weak criterion drags the module: strong ideas cannot fully compensate for grammar that interferes with meaning, and fluent delivery cannot buy back an unanswered question.

Using the arithmetic to plan your preparation

Once the machinery is visible, preparation becomes a resource-allocation problem. Suppose you need overall 7.0 with no module below 6.5, and your mock scores are Listening 7.5, Reading 6.5, Writing 6.0, Speaking 6.5. The average is 6.625 — reported as 6.5 — so the headline gap looks small. But the real blocker is Writing at 6.0, which breaches the minimum-module rule regardless of the average. Raising Reading from 6.5 to 7.0 costs perhaps four extra correct answers and lifts the average; raising Writing from 6.0 to 6.5 requires shifting at least one examiner criterion, which is slower work. The efficient plan does both, but it books the test date around the Writing timeline, because that is the score that cannot be rushed. Candidates who plan around the overall number alone routinely resit with the same module profile.

• Convert every practice result into module bands before deciding what to study next — raw totals hide where the gap is.
• In Listening and Reading, count the marginal questions: half a band is usually three to four answers, and your error log says which type they come from.
• In Writing and Speaking, target the weakest criterion by name; averaging means one criterion at 6 caps what the other three can deliver.
• Recheck the minimum-module condition of your target organisation — it, not the overall band, is usually the binding constraint.

The overall band: averaging and rounding

The overall band is the mean of the four module bands, rounded to the nearest half band — and the rounding is generous: an average ending in .25 rounds up to the next half band, and one ending in .75 rounds up to the next whole band. So Listening 6.5, Reading 6.5, Writing 6.0 and Speaking 7.0 average 6.5 exactly; swap the Writing 6.0 for 6.5 and the 6.625 average still reports as 6.5. This is why requirements phrased as 'overall 7.0 with no module below 6.5' are stricter than they look — the minimum-module condition is usually the binding one.