How celpip9 estimates your level, and what the estimate cannot tell you

What a celpip9 level is, and what it is not

Every level celpip9 shows is an estimate on the Canadian Language Benchmarks (CLB) scale. It is always shown as a range around the estimate, never as a single point, and never as a CELPIP score or result. A CELPIP result comes from the test provider after test day and is reported on levels 3–12, where each level corresponds to the CLB level of the same name, as the test provider explains on its scoring page. We use the CLB scale so you can read your practice in familiar terms. The range does not come from the test provider, and it carries none of the weight of a test result.

celpip9 is an independent practice platform. It is not connected to, approved by or endorsed by the test provider or by IRCC, as the footer of every page says. Why show a range? There is some uncertainty at each step between your answers and a level, and a single number would hide it. A band that spans, for example, CLB 8 and CLB 9 shows that uncertainty openly. A single level would suggest a precision we cannot yet show.

How Listening and Reading estimates are worked out

For Listening and Reading, the estimate starts with a simple count of the questions you answered correctly. We read that count through the approximate chart the test provider publishes, which links numbers of correct answers to levels. The chart is approximate, and it was built for the test's own questions, not for ours.

Our practice tests are not yet equated to the scale the test itself uses. Equating is the statistical work that makes different test forms comparable, so that the same number of correct answers means the same thing whichever form a person takes. Without it, a practice set of ours could be slightly easier or harder than a test-day form, so the same count could point to a different level. That is part of the reason the estimate is shown as a range rather than a point.

Our practice items are written by AI from the published format, never from the test provider's materials. We use an item only after independent checks agree on it. The checks are there to catch items that are wrong or unclear. They do not make a practice set statistically equal to a test-day form, so they do not remove the need for a range. Our methodology describes both steps.

Where Writing and Speaking scoring stands today

Writing and Speaking are not scored on the site yet. If you complete a Writing or Speaking task now, we keep your response in your report but do not give it a level. The status page shows what is live today, so check it before relying on any Writing or Speaking estimate.

Here is how the scoring we are building works. Each response is judged against the dimension names the test provider publishes in its performance standards for Writing and Speaking. The judging happens in several independent passes. When passes disagree by a level or more, we add more passes. We then report the middle result, shown as a range like every other celpip9 level.

Judging against published dimension names does not reproduce the test provider's own scoring. Taking the middle of several passes reduces the effect of one unusual judgement, but some uncertainty remains. We also publish no templates, openers or phrases to learn by heart. The test provider warns that responses that are not original can put a result at risk, for example by being cancelled. Our guide explains why your own words, organized for the task, are the safer path.

What we measure about ourselves, and what we have not measured yet

The trust panel publishes what celpip9 measures about its own work, such as answer-key accuracy (whether the keyed answers to practice items are correct) and scoring consistency (whether repeated judgements of the same work agree). Each figure is shown with its sample size and the date it was measured, so you can see how much evidence supports it. We do not repeat those figures here because they change. The panel always has the current ones.

Anything we have not measured is marked not yet measured on the panel. The most important of these is how well our estimates match the levels test takers later report after test day. Until that is measured, no one, including us, can say how often a celpip9 range contains the result a person goes on to receive. We make no claim about it. Any claim about predictive accuracy should come with evidence you can check.

What the estimate cannot tell you

It cannot predict your result. A range shows how your practice work reads against a published chart, or against published dimensions once Writing and Speaking scoring goes live. It is not a forecast of test day, where the questions, the setting and how you feel on the day will all be different.

It cannot replace a result that an institution accepts. A practice estimate has no standing in an application. For what IRCC requires, IRCC's own pages are the authority: see its pages on language test results for Express Entry and language proof for citizenship. Rules change and have exceptions. This article is general information, not immigration advice. For an overview of what IRCC asks for, see CLB levels for Express Entry and citizenship.

It cannot describe what a level means in terms of ability. A range that reaches CLB 7 tells you where your practice work sits on a scale. It does not say what a person at that level can do at work or in daily life. For how CELPIP results relate to CLB levels, see how CELPIP results work.

How to use a range in your preparation

Use the range as a direction, not a verdict. Look at where the range sits across several practice sets rather than reacting to any single set. If your Reading range stays lower than your Listening range over time, that is a fair sign of where to spend more effort, and the component guides can help you focus. A range that moves by a small amount between sets is usually within the uncertainty the range is there to show.

Before you practise under time, the test drive lets you try every mechanic of the test: audio that plays only once, the live word count and the recorder that starts and stops on its own. It is untimed and unscored, so you come away familiar with the mechanics but without a level. To plan your weeks, see how to structure your CELPIP preparation. To find the source of any figure on this site, see sources.