Outcomes
How we measure what we claim
Any score figure we publish is computed from our own students' assessment records under the rule set out below. We show the number of students it is based on, every time. Where we do not have enough students to say something meaningful, we say nothing rather than something smaller.
What we can currently say
We couldn't load our figures just now. This is on our end, not a statement about what we have measured — please try again shortly. The rule below is what any figure we publish has to meet, and it does not change.
We would rather say this than leave you looking at a blank space, or at a number we did not actually read.
The rule
Who is counted
A student is included when all of the following are true:
- • They have at least two recorded assessments for the same exam. We compare their first to their last — never their worst to their best.
- • The two sittings are at least 30 days apart. Two tests a week apart measure a good day, not teaching.
- • They attended at least 4 coaching sessions between the two. A student who bought a diagnostic and never came back improved on their own; that is not ours to claim.
- • The first sitting is within the last 2 years, so the figure describes the programme as it is taught now.
Who is not
- • Students with one assessment. They are excluded, not counted as zero — being measured once is not the same as not improving, and treating it as zero would understate the figure as dishonestly as dropping poor results would overstate it.
- • Nobody is excluded for going down. Students whose score fell are in the average. A number computed only from students who improved is not an average.
- • Cancelled and no-show sessions do not count towards the session minimum.
When we publish nothing
Below 25 students we publish no figure at all — not a smaller one, not a wider range, not "based on a small sample". A number that small will be quoted back without its caveat. The same threshold applies to a single cohort as to the headline; small cohorts are combined rather than published.
What this does not prove
- • This is not a controlled trial. There is no comparison group of similar students who did not work with us. Some of the improvement would have happened anyway — students retaking a test tend to score higher the second time, with or without coaching.
- • The students are self-selected. Families who buy coaching and attend it are not a random sample of test-takers.
- • Scores are as recorded. They come from what students and their coaches entered, including practice tests, not only from official score reports.
- • It says nothing about admissions. We do not publish an admissions figure because we cannot attribute one.
Figures are recomputed daily from stored records. If you are a current family and your dashboard shows a different number from this page, that is a bug worth telling us about — both are produced by the same calculation.