Our level test asks you about ten questions, not forty-four. It is not a cut-down version: it is a test that picks each question from your previous answer. Here is how it actually works, what it does not measure, and the point where the decision passes to a person.
Where it starts
We do not ask you what we already know you know
In a fixed test, with the questions ordered from the easiest to the hardest, most of them are settled before you start: if your English is good you will almost certainly get all the elementary ones right, and get almost all the very advanced ones wrong. Those questions cost you time and tell nobody anything.
A question is only worth asking when nobody knows how it will turn out. Pick a level below and look at the difference between the two ways of doing the same thing.
How it works
Long strides to get there, small steps to close in
The test keeps a single estimate of your level, starting from the middle of the scale. Answer well and it goes up one step; get it wrong and it goes down. The next question is always taken at the level the estimate has reached: that is why the questions seem to change difficulty as you go.
The interesting part is the step. Every time you change direction — a right answer after a wrong one, or the other way round — the step gets shorter. First you stride over the edge, then you feel for it with small steps. Once the step is short enough, going on would no longer change the result: the test stops there.
See the answers one by one
| # | Question at level | Outcome | Chance of getting it right | Step | Estimate after |
|---|
The margin
Three questions would be enough, if nobody ever slipped
That is why the test stops at a minimum of ten questions and a maximum of fifteen. Going further would ask more of your time in exchange for a precision that five minutes of conversation give better.
The decision
When the result is not clear-cut, a person decides
Along with the level, the test also produces how reliable that level is, that is, how far the step narrowed by the end. It is not a technical detail: it is what decides what you see and what lands on a teacher’s desk.
If the result is clear-cut
You see your level and the course that matches it. A teacher from our team reviews your answers anyway and writes to you within one working day.
If the result is not clear-cut
Non ti diamo un livello secco: vedi un intervallo di due livelli, detto com’è. Le tue risposte passano a un docente, che decide lui e ti scrive la sua valutazione entro un giorno lavorativo.
In both cases the path starts with fifteen minutes of conversation with a teacher, free, and that is where the level is confirmed. The test is there so nobody wastes time in that conversation, not to replace it.
Said upfront
What these ten questions do not measure
- They stop at C1. Se rispondi bene anche alle più difficili che abbiamo, il test non ti sta misurando: ha finito la scala. In quel caso il risultato dice “C1 o oltre”, e quanto oltre si capisce parlando.
- At the two extremes the estimate can be off by one level. If you are well above or well below the average you meet few questions at your own height, because the scale ends. That is why the level always travels with its reliability.
- They measure grammar and vocabulary in writing. Pronuncia, scioltezza e capacità di reggere una riunione non entrano in una domanda a scelta multipla: si vedono nei quindici minuti di conversazione, e poi nell’assessment finale del percorso, valutato da un docente.
- It is not a certification. It is the starting point for picking the right level, and for not making you start over if you are already ahead.
It is free and takes five minutes. Would you rather talk it through first? Book 15 minutes with a teacher.
