There is no examiner
This is the single most useful thing to know before you prepare, and most people find out on test day.
CELPIP is delivered entirely by computer — Speaking included. You sit at a screen, a prompt appears, and you speak into a microphone. Your answers are recorded. Nobody is sitting across the table from you.
That changes how you should prepare, in three specific ways:
- There is no rapport to build. No smiling, no eye contact, no reading the examiner's face to see whether you are doing well. If you have practised for an interview, you have practised for a different test.
- Nobody will help you out. A human examiner might nod you along, or repeat a question. The screen will not. If you stop talking, silence is what gets recorded.
- The clock does not care. It starts and it runs. There is no "sorry, could I start again?"
The practical consequence: practise talking to a device with nobody in the room. It feels strange, and that strangeness is exactly what you want to get out of the way before test day rather than on it.
The arithmetic nobody puts in front of you
Paragon publishes 15 minutes for the Speaking section, and 8 tasks.
Divide it. Under two minutes per task — and that includes your preparation time.
That single number reorganises your whole preparation. You are not being asked to give a considered, polished answer. You are being asked to start talking almost immediately and keep going in a straight line. A person who needs thirty seconds to decide what to say has spent a quarter of the task deciding.
Paragon does publish the per-task split, though not on the test-format page — it is in the Speaking study packs:
| Task | Preparation | Speaking |
|---|---|---|
| 1 Giving Advice | 30 s | 90 s |
| 2 Personal Experience | 30 s | 60 s |
| 3 Describing a Scene | 30 s | 60 s |
| 4 Making Predictions | 30 s | 60 s |
| 5 Comparing and Persuading | 60 s, then 60 s | no speaking in part 1, then 60 s |
| 6 Difficult Situation | 60 s | 60 s |
| 7 Expressing Opinions | 30 s | 90 s |
| 8 Unusual Situation | 30 s | 60 s |
Read the preparation column again. Thirty seconds, on six of the eight tasks. That is the whole preparation you get — not two minutes, not one. Tasks 5 and 6 give you a minute, and Task 5 is the odd one: it runs in two parts, and the first part is preparation with no speaking at all.
All eight tasks
| # | Task | What it asks | Where it is lost |
|---|---|---|---|
| 1 | Giving Advice | Advise a specific person about a specific problem | Generic advice that would suit anyone |
| 2 | Talking About a Personal Experience | Tell about something that happened to you | A story with no point — events, no meaning |
| 3 | Describing a Scene | Describe an image | Listing objects instead of describing a situation |
| 4 | Making Predictions | Say what will happen next in that scene | Describing again instead of predicting |
| 5 | Comparing and Persuading | Compare two options, then persuade someone | Comparing and never persuading |
| 6 | Dealing with a Difficult Situation | Choose how to handle an awkward situation and explain | Avoiding the choice |
| 7 | Expressing Opinions | State an opinion and support it | Stating it and then repeating it |
| 8 | Describing an Unusual Situation | Describe an odd image to someone who cannot see it | Forgetting the listener is blind to it |
Read that last column again. Six of the eight are lost by answering a slightly different question from the one asked. Not by weak grammar. Not by accent. By drifting one step off the task.
What is actually marked
Each CELPIP skill is scored 0 to 12, and that maps directly to a CLB level. Paragon publishes the four Speaking dimensions by name: Content/Coherence, Vocabulary, Listenability, and Task Fulfillment. Listenability is defined as "How easy is it to listen to and understand the response?"
Your accent is not one of them, and Paragon says so directly:
"Our raters specifically do not consider your accent when rating your responses."
Millions of people speak clear, fluent English in an Indian, Nigerian, Filipino or Pakistani accent, and CELPIP is not marking that.
One honest qualification, because it is the thing people confuse. Accent is not rated — pronunciation is. Listenability's own sub-points are "Rhythm, pronunciation, and intonation", and the level descriptors talk about whether "pronunciation errors rarely impede communication". The line is being understood, not sounding Canadian. Nobody is marking where you are from; somebody is marking whether a listener has to work.
And the rule that outranks everything else:
A response giving insufficient information to assess scores 0 or 1.
Talking around the task for ninety seconds without answering it is not a low score. It is the bottom of the scale.
Three habits that carry across all eight
- Start with the answer. Not "well, that's an interesting situation…". Say the advice, name the choice, state the opinion — in the first sentence. Then spend the rest supporting it.
- Keep going. A slightly clumsy sentence that continues beats a perfect sentence followed by four seconds of silence. There is no examiner to fill the gap.
- Answer the task you were given. Re-read that "where it is lost" column. It is the same failure eight times over, wearing different clothes.
Try one
[Try a Speaking task →] — one full task, marked, no card required.