A practice result and an exam result answer different questions. This page says, briefly, how the CELPIP exam is scored, then what you get after practice on celpiptest.
How the exam is scored
Listening and Reading are scored by computer. Each question is right or wrong. The raw result is then equated and turned into a CELPIP level. The writing rubric is not used here.
Writing and Speaking are scored by trained raters. They use four dimensions: Content/Coherence, Vocabulary, Readability or Listenability, and Task Fulfillment. Several raters score independently. For the full account, read how each section is scored.
After Writing or Speaking practice
When you finish a task in Writing or Speaking, you see an estimated level. You see which criteria the estimate was based on: the same four. You see why the estimate landed there. You also get a recommendation for what to change to improve the result.
The estimate is produced against the published criteria and then checked against the task rules. It is not a single impression of fluency or grammar. The response is read against each of the four criteria. It is then checked against the task: whether the required points are covered, whether the answer is long enough, and whether it stays on the prompt. A fluent or tidy answer that misses the task does not get to float upward. Those checks are there so the estimate stays as close as we can make it to the way the exam is scored. It is still an estimate, not a CELPIP score report.
Listening and Reading practice
Listening and Reading practice are scored from the answer key. Each question is right or wrong, and the result is not based on those four dimensions.