How is your IELTS Writing band score calculated?

01. The four criteria, in plain English

Both Task 1 and Task 2 are scored on the same four criteria, defined in the official IELTS scoring guide. Each criterion is weighted equally — 25% — toward the Writing band for that task. The four are:

• Task Response (TR) — Did you actually answer the question? Did you cover every part of the prompt? Did you give a clear position and develop it with relevant ideas? For Task 1, this is renamed Task Achievement and measures whether you reported the key trends or stages accurately.
• Coherence and Cohesion (CC) — Can the reader follow your argument without going back? Are paragraphs organised around one main idea each? Are linkers ("however", "as a result", "in contrast") used naturally, not sprinkled in for show?
• Lexical Resource (LR) — Range and accuracy of your vocabulary. Do you use a variety of words appropriate to the topic? Are spelling and word formation accurate? Do collocations sound natural ("strong argument," not "powerful argument")?
• Grammatical Range and Accuracy (GRA) — Variety of sentence structures (simple, compound, complex) and correctness. Punctuation, agreement, tenses, articles. Errors that obscure meaning cost more than errors that don't.

02. What each band actually requires

The IELTS public band descriptors define what a Band 5, 6, 7, 8 or 9 looks like for each criterion. They are not secret — IELTS publishes them, and examiners use them to make consistent judgements. A few practical takeaways most candidates miss:

On Lexical Resource, Band 7 wants "less common lexical items" and an "awareness of style and collocation." Throwing in a thesaurus word you don't usually use will often hurt your score, not help it — because it tends to break collocation or register.

03. Task 1 vs Task 2 weighting

Your overall Writing band is a weighted average: Task 2 counts twice as much as Task 1. A 7 on Task 2 and a 6 on Task 1 averages closer to 6.5 than to 7. This is why most candidates should spend roughly 40 minutes on Task 2 and 20 on Task 1, and why a great Task 1 cannot rescue a thin Task 2.

Within each task, the four criteria are equally weighted. The Task 2 band is the average of TR, CC, LR, and GRA, rounded to the nearest half-band. So a single weak criterion drags the whole task — improving from a 6 to a 7 in your weakest area lifts the overall task band by 0.25 on its own.

04. From criteria scores to your overall band

The full path looks like this:

1. The examiner reads your response and rates each of the four criteria from 0 to 9 against the band descriptors. Half-bands are allowed.
2. The four scores are averaged to give a Task band, again rounded to the nearest half.
3. The two task bands are combined with Task 2 weighted ×2, then rounded to the nearest half. So Task 1: 6.5 and Task 2: 7 → (6.5 + 7 + 7) / 3 = 6.83 → final Writing 7.0.

Knowing the math matters because it tells you where the highest-leverage improvements are. If you're at 6.5 because of LR specifically, you don't need to rebuild everything — you need a targeted plan for vocabulary range and spelling.

05. The four most common misconceptions

• "Spelling errors will tank my LR." Spelling is one signal in Lexical Resource. A few minor errors at Band 7 are tolerated. Repeated basic spelling errors push you below Band 6.
• "I just need to write more words to get a higher band." The word counts (150 for Task 1, 250 for Task 2) are floors, not targets. Going well over without adding value risks lowering coherence and lexical accuracy.
• "Memorised templates will lift my Task Response." They typically hurt it. Examiners are trained to spot template language and will discount it from your assessment.
• "My grammar is fine because I make no big mistakes." Avoidance is not the same as range. If you only write what you know is safe, your GRA will plateau at Band 6 to 6.5.

06. How to use this for practice

Most candidates self-assess in vague terms ("that essay felt good"). Switch to criterion-by-criterion. After each timed essay, write a quick post-mortem:

When you ask for an AI band estimate, you'll get exactly this breakdown — a score per criterion plus the specific issues that pulled each one down. The goal isn't the score itself; it's the next concrete thing to fix.