How the paper is built
You get sixty minutes for two tasks and you manage that time yourself — no one stops you at the twenty-minute mark. The recommended split is twenty minutes for Task 1 and forty for Task 2, and it is recommended for a reason: Task 2 carries twice the weight of Task 1 in the writing score. Spending half your time on the task worth a third of the marks is one of the most common self-inflicted wounds on the paper.
| Academic | General Training | |
|---|---|---|
| Task 1 | Describe visual information — a line graph, bar chart, pie chart, table, map or process diagram — in at least 150 words. You report what the data shows. You do not explain why or give an opinion. | Write a letter of at least 150 words in response to a situation. The prompt sets the recipient, which sets the register: formal, semi-formal or personal. |
| Task 2 | A discursive essay of at least 250 words responding to a position, an argument or a problem. Identical in form to the General Training essay, but the prompts tend to be more abstract. | A discursive essay of at least 250 words, usually on a more everyday subject. |
| Suggested time | 20 minutes for Task 1, 40 for Task 2 | 20 minutes for Task 1, 40 for Task 2 |
| Weighting | Task 2 counts for twice as much as Task 1 | Task 2 counts for twice as much as Task 1 |
The two tasks, in both versions of the test.
Both tasks are marked by a trained examiner against four published criteria, each contributing a quarter of the score for that task. You are given a band for each criterion, and the task band is their average — so a very strong answer with one weak criterion cannot score at the top.
The four criteria, in plain English
The official descriptors are detailed and public, and reading them is worthwhile. What follows is not a reproduction of them but an explanation of what each criterion rewards in practice, and what candidates most often get wrong under each heading.
Task Achievement (Task 1) / Task Response (Task 2)
Did you do what was asked, completely? In Task 1 that means covering the main features of the data with accurate figures and an overview. In Task 2 it means answering every part of the question — and a great many prompts contain two parts, of which candidates answer one.
The commonest cause of a low band here is a good essay on a slightly different question.
Coherence and Cohesion
Can the reader follow the argument without effort? This is paragraphing, logical order and the connections between sentences — and it is where mechanical linking words do damage. Firstly, Moreover, In conclusion stapled onto disconnected ideas reads worse than no connectors at all.
One clear idea per paragraph, stated in the first sentence.
Lexical Resource
Range and precision of vocabulary, and whether it is used naturally. Precision beats rarity: a common word used exactly is worth more than an unusual one used approximately, and a misused sophisticated word costs more than a plain accurate one.
Collocation errors are more visible to an examiner than simple vocabulary.
Grammatical Range and Accuracy
Both halves matter, and candidates usually optimise for one. Simple sentences that are all correct will not reach the higher bands; complex sentences riddled with errors will not either. The target is complex structures that stay accurate.
Article and preposition errors are the most persistent at B2–C1.
Notice what is absent from that list. There are no marks for your opinion being agreeable, for the essay being interesting, or for knowing anything about the topic. The examiner is assessing English, and a well-argued position they disagree with scores exactly as well as one they share.
The penalties that quietly cap a score
Several rules are applied mechanically, and each one is capable of holding down an answer that is otherwise perfectly competent. They are easy to avoid once you know they exist, which is the only reason to list them.
- Under-length. Below 150 words in Task 1 or 250 in Task 2, the score is penalised. Write comfortably over — around 170 and 280 — because counting words in an exam wastes time you do not have.
- Off-topic writing. An answer that does not address the question set is heavily penalised, however good the English. This includes a memorised essay bent towards the prompt, which examiners are trained to recognise.
- Copying the prompt. Words lifted from the question do not count towards your total, so an introduction that reproduces the prompt is both shorter than it looks and marked down for it.
- Bullet points and note form. Both tasks require continuous prose. Notes, lists and headings are penalised under Coherence and Cohesion.
- Opinion in Academic Task 1. The task is to report the data. Explaining why a trend occurred, or arguing about it, is outside the task and does not earn credit.
- Illegible handwriting on the paper test. What cannot be read cannot be marked. If your handwriting is poor under speed, the computer-delivered test removes the risk entirely.
Academic Task 1: describing data without describing everything
The instinct is to list every number in the chart. That produces a low band under Task Achievement, because the task is to select and report the main features — the shape of the data, not its contents. A four-paragraph shape handles almost every version of the task.
- One sentence of introduction. Say what the chart shows, in your own words rather than the prompt’s.
- An overview. Two sentences on the biggest patterns: what rose, what fell, what was largest and smallest overall. An answer with no overview cannot reach the higher bands, and this is the most commonly missing paragraph on the paper.
- Two body paragraphs. Group the data logically — by trend, by category, or by period — and support each point with a specific figure. Comparison is what earns marks; sequential listing is what loses them.
- No conclusion. Task 1 has an overview instead. A conclusion repeating it adds words and no marks.
The language of trends
Precision here is easy marks, because the vocabulary is small, closed and reusable. Vary the grammar rather than hunting for exotic synonyms.
- Car ownership rose sharply between 1990 and 2005, from 12 million to just over 30 million.
- There was a steady decline in rail use over the same period.
- Bus journeys fluctuated throughout the 1990s before levelling off at around 8 million.
- Cycling accounted for the smallest proportion of journeys in every year shown.
- The figure for walking more than doubled, whereas that for motorcycling remained largely unchanged.
Two grammatical points carry a surprising amount of weight. Tense follows the data: a chart with dates in the past takes past tenses, a projection takes future forms, and a process diagram usually takes the present simple passive — the beans are roasted, then ground. And figures need the right preposition: rose from X to Y, an increase of 12%, an increase in production, at around 8 million.
Task 2: the essay that carries two-thirds of the paper
Task 2 asks you to respond to a position, an argument or a problem in at least 250 words. Prompts recur in a small number of shapes, and identifying the shape correctly is most of Task Response.
Opinion — “To what extent do you agree?”
Take a clear position in the introduction and hold it throughout. Balanced-to-the-point-of-invisible essays lose marks here, because the task asked for a view and the answer did not supply one.
A qualified position is fine. An absent one is not.
Discussion — “Discuss both views and give your opinion.”
Two parts, and both are compulsory. Give each view genuine space, then state your own — an essay that discusses both and forgets the opinion has answered half the question.
One paragraph per view, then a short opinion paragraph.
Problem and solution
Identify causes or problems, then propose specific responses. Vague solutions — the government should take action — are the weak point; name the action and say who takes it.
Two developed solutions beat five listed ones.
Advantages and disadvantages
Read the second half of the prompt carefully: it often asks whether the advantages outweigh the disadvantages, which requires a judgement, not a list of both.
If it asks you to weigh them, weigh them explicitly.
A structure that works for all four
- Introduction, two sentences. Paraphrase the topic, then state your position or say what the essay will do. No background, no rhetorical questions, no statistics you invented.
- Body paragraph one. One idea, stated in the first sentence, then explained, then illustrated with a concrete example. The explanation is the part candidates skip and the part that earns the band.
- Body paragraph two. A second idea, developed the same way. Two well-developed ideas outscore four undeveloped ones every time.
- Conclusion, two sentences. Restate the position in different words and close. No new arguments, no apology, no plea to the reader.
Improving writing when nobody is marking it
Writing is the hardest paper to self-prepare, because the thing you most need — accurate feedback — is the thing you cannot give yourself. There are still four habits that work without a teacher.
Write it twice
Write the essay in forty minutes. The next day, before rereading it, write the same essay again with no time limit. The differences between the two are precisely your list of pressure-induced errors, and it is a short, personal and highly actionable list.
Cheaper and sharper than most feedback.
Keep an error log
Record every mistake you find in your own writing, grouped by type rather than by essay. Most candidates discover they are making four or five errors repeatedly — usually articles, prepositions, subject–verb agreement and tense consistency.
Study cohesion as grammar
Coherence marks come from structure, not from linking words. Relative clauses, participle phrases and controlled use of the passive do more to connect ideas than a list of connectives ever will.
Build topic vocabulary, not phrase banks
Task 2 prompts recycle a small set of subjects. Learning the vocabulary of those fields properly gives you something to say and something to say it with, which memorised phrases never do.