Do learners get better?
A school should show its results, not only promise them. This page publishes ours, with the method and how many learners stand behind each number.
What is measured, and how
Two figures, each taken in a learner's first week (days 1 to 7 after their first activity) and fourth week (days 22 to 28):
- Calibration: the Brier score of every answer given with a confidence (60, 75 or 90% sure). It is the squared gap between the confidence and what happened: 0 is perfect, and always saying 60% on answers that are right half the time scores 0.26. Lower is better.
- Process: the average reading score, out of 7, on simulator charts. Every chart is new to the learner, so the score is measured on charts the learner has never practised on.
A week counts only with at least 10 answers with a confidence and 3 chart readings. Results in R are left out on purpose: on these practice charts they are mostly luck. The figures come only from learners who choose to send them, which can flatter the result, so the number of learners is always shown next to each figure.
Results so far
Loading.
Your figures
Not enough yet. Answer with a confidence in the simulator and the drills for a week, and your week 1 figures appear here.
Only the four numbers in the table above are sent, with a random id for this browser so a week is not counted twice. Nothing else leaves the browser, and you can switch this off at any time.