What a TEF Canada sample test actually tells you, and how to read your own result

Most candidates sit a sample test hoping for a number, get one, and then treat it as a forecast. It is worth being blunt about this: a sample test is a format document first and a measuring instrument second, and it measures the four abilities with very different degrees of honesty. Reading it as a single predicted score throws away almost everything useful in it.

The value of a sample is not the total. It is the pattern underneath the total, and that pattern only becomes visible if you record the result in the right shape while you are marking it.

What a sample is for

A sample exists to show you what the questions look like and how they are asked. That is a real service. The TEF Canada asks you to do four separate things, and each of them has a house style: the way the listening items are worded, the length of the reading texts, what the writing tasks hand you as a starting point, what the speaking examiner is waiting for you to produce.

Working through a sample turns those from unknowns into routine. On the day, that is worth several minutes of not being confused, and several minutes is a lot.

What a sample is not for is telling you your level from one sitting. One paper has too few items to be stable, and the productive sections have no answer key at all.

Reading the receptive half of your result

Compréhension orale and compréhension écrite are the two abilities where a sample gives you something close to a real measurement, because the answers are objective. Mark them strictly, with no partial credit for nearly, and record two numbers rather than one: how many you got right, and how many you never reached.

Those two numbers are separate diagnoses. A candidate who answers everything and gets a third wrong has a language problem. A candidate who gets almost everything right but leaves the last several blank has a pacing problem, and pacing is much faster to fix. Collapsing both into a single percentage hides the difference, which is why a bare score is such a poor thing to carry around.

Then convert. Raw marks mean nothing until they are read against the published equivalency, and each ability converts on its own. Use the TEF Canada score chart and take your level from the weakest of the four, because that is how your file will be read. Averaging four abilities is the most common way candidates talk themselves into a level they do not have.

Reading the productive half, where the sample gives you least

Here the sample hands you a prompt and stops. Nothing in it can tell you what a marker would have given you, and self-assessment on writing and speaking is unreliable in a predictable direction: candidates who are weak on range judge themselves on accuracy, because accuracy is the thing they can check.

You can still extract something honest from it if you assess process rather than quality. After a sample writing task, answer four questions about what you produced.

• Did I finish inside the time, or did I run over?
• Did I write to the length the task asked for?
• Did the text do what the task asked, or did it drift into a related topic I was more comfortable with?
• How many subordinate clauses did I actually write, and were they all the same kind?

None of that is a score. All of it is checkable by you, and the fourth question in particular predicts a band better than most candidates expect, because range is what separates a safe middle text from an upper one. What the marking grid is looking for on the TEF writing section is not an absence of errors.

For expression orale, record yourself and listen back a day later. The delay matters. Listening immediately, you hear what you meant to say; a day later you hear what came out, including the pauses and the three fillers you use without noticing.

Turning one sample into a diagnostic

A single sample is a snapshot with a lot of noise in it. Two samples sat under the same conditions two weeks apart are a direction, and a direction is far more useful than a number.

Keep the conditions fixed between them: same time of day, same order of sections, same absence of pauses, same lack of a dictionary. If you change the conditions you have measured nothing, because you cannot tell whether the difference came from your French or from the fact that you sat the second one on a quiet Sunday morning.

If the two results are flat, look at the cause breakdown rather than the total. Flat with a shrinking number of unreached questions is progress that the score has not caught up with yet. Flat with the same errors in the same places is a plateau, and it means the material is rehearsing what you already know.

Choosing what to do with the result

Rank your four converted levels and look only at the lowest. That ability is the one deciding your file, and time spent on any of the other three changes your points without changing your eligibility. If you are not sure what the threshold means in practice, the NCLC 7 page sets it out.

If your two lowest abilities are the productive ones, which is the common case, more samples will not move them. If you are still deciding between the two exams at this stage, the format differences are laid out in TEF Canada vs TCF Canada, and it is worth settling that before you invest in sample material for one of them.

One last point of caution. Sample material circulates widely and much of it is out of date or was never official. Confirm the current structure and timings against the CCI Paris Île-de-France, which runs the exam. Français Prep is an independent preparation service and is not affiliated with it.