CELPIP Writing: Two Tasks, 53 Minutes, and One Rule That Outranks Your English

The rule that outranks your English

53 minutes, two tasks — and the thing that decides your band is not your grammar, it is whether you answered everything you were asked. For Writing and Speaking, the harshest official rule applies: a response giving insufficient information to assess scores 0 or 1.

Not marked down. Marked out. A beautifully written email that misses one of the prompt's three bullets is not an 8 with a deduction; it is at risk of the bottom of the scale. And 0 and 1 are not ordinary low scores — on CELPIP's chart they carry that same label and map to no Canadian Language Benchmark level at all.

The failure: writing well about two-thirds of the prompt

One detail in the situation catches your interest, and you write four strong paragraphs about it. Tenses clean, register right, time to spare. You covered two of the three things the prompt listed.

This is a fluent writer's mistake. The better your English, the easier it is to fill 180 words with well-made sentences about the part you found easiest, and never notice the bullet you skipped. The prompt is not a topic. It is a checklist.

The two tasks

Each task has its own page, linked below. What follows applies to both.

The arithmetic — and it is only arithmetic

Writing is 53 minutes for 2 tasks. Divide, and you get roughly 26 minutes each, including reading the prompt, planning, writing and checking.

Say plainly what that number is: it is division, not a rule. No official per-task split is published — 53 minutes is the section total. Treat 26 minutes as a figure you impose on yourself, not a deadline the test announces. It still earns its keep: 26 minutes is not much for 150 to 200 words if you spend eight of them deciding what to say, and plenty if you spend two.

The counting habit

The highest-value thirty seconds in the section is not a writing skill at all. Count the demands in the prompt before you write. Answer that many. Then check you answered that many. Concretely:

1. Read the prompt with a finger on the bullets. Paragon's Task 1 sample has three. Say the number in your head: three.
2. Write the number down, then the three demands beside it as three words. Not sentences — three words. Dates. Refund. Contact.
3. Give each one its own paragraph, or at minimum its own sentence. A demand buried inside a paragraph about a different demand is easy for a rater to miss — and if it can be missed, it has not been credited.
4. At the end, do not reread for grammar first. Count. Point at each word and find the sentence that answers it. A word with no sentence is the thing that would have cost you most, found with time left to fix it.

Drill this on every practice attempt — on test day the pressure pushes you toward writing, not checking, and the habit has to be older than the pressure.

Task 2's demands are not always bullets: choosing an option and explaining it is two demands, and the second is where people run out of material.

Why "about" is doing real work in "about 150-200 words"

Paragon's sample instruction says to write "about 150-200 words". "About" is guidance on scale, not a threshold.

An email of 211 words is not penalised for being 211. The range tells you the size of response the task is built for — long enough to cover three demands properly, short enough not to be an essay.

Short responses do score badly, but the number is a symptom, not the cause. A 90-word email is usually short because it left something out — and that is what costs you. If your draft lands at 90 words, do not pad the closing; find the demand you skipped.

The tools on the screen

The test is fully computer delivered and "features such as a personal timer, word counter, and spell-check", on a standard QWERTY English keyboard.

The word counter is useful exactly once: at the end, to confirm you are in the right region. One glance, after the counting check.

It is not useful while you compose. Writing to a number turns a covered demand into a padded sentence and an uncovered demand into a problem you never look for — you finish at 175 words feeling finished, which is the feeling the counting habit exists to interrupt. Set the personal timer against your own 26-minute figure, and save spell-check for the final pass, after the count.

What no page can tell you

Listening and Reading have answer keys — each is reported against a raw score out of 38, and a question is right or it is not.

Writing has no key. Both tasks are prose, judged rather than matched to a correct answer, so no page — this one included — can tell you the exact CELPIP level a given response would score. Anyone showing you a sample captioned "this is a 9" is offering an opinion, not a measurement.

That is a reason to change target, not to despair. You cannot verify your band from your desk, but you can verify on every attempt that you answered every demand in the prompt. Drill the countable thing.

If you think a Writing score is wrong, CELPIP does allow a re-evaluation — a re-mark, not a retake — within six months, one per component. Check the current terms on its policies page.

What actually costs you

• Covering two of the three bullets, fluently.
• Rereading for grammar at the end and never rereading for coverage.
• Burying a second demand inside a paragraph about the first, where a rater can miss it.
• Padding a thin response to hit a word count instead of finding the missing content.
• Watching the word counter while you write.
• Spending eight minutes choosing a position in Task 2 when either option can score at the top.
• Letting Task 1 eat Task 2 because you never set your own per-task figure.

Try one

[Write a Task 1 email →] — one full task, marked, no card required.