How We Will Rank Schools, and Why We Have Not Yet
Publishing the method before the table, so it can be argued with before it can be used. What we will measure, what we refuse to measure, and why most school rankings tell you about intake rather than about teaching.
The single most valuable thing this site could publish for a relocating family is a ranking of the international schools within reach of this coastline. We have not published one, and this piece explains what we will do instead, in the order we will do it, so that the method can be attacked before it produces a table rather than afterwards.
The problem with the tables that exist
Nearly every school ranking is built on outcomes — exam results, university destinations — and outcomes at selective schools are substantially determined by admission. A school that admits able children from households with books in them and a parent who reads reports will produce able results, and will have demonstrated very little about its teaching.
That is not an argument that results are meaningless. It is an argument that a raw results table ranks intake, and that a family choosing between four schools with similar intakes learns almost nothing from it.
A league table of selective schools mostly measures who was let in. It is a ranking of admissions offices wearing the clothes of a ranking of teachers.
What we will measure
- Curriculum and continuity — which programmes are offered, at which stages, and whether a child can go from entry to university without switching system. Switching curriculum mid-secondary is the most damaging avoidable thing a relocating family does.
- Capacity and admission reality — actual year-group sizes, whether a waiting list operates, and at which points in the year a place can realistically be taken. A school that is excellent and full is not available to you.
- Staff stability — teacher turnover and average tenure. This is the closest available proxy for what a school is like to work in, and it correlates with what it is like to attend.
- Provision for children arriving mid-system — language support, and what happens to a twelve-year-old who lands in March from another country. Every relocating family needs this and almost no prospectus addresses it.
- The commute, honestly stated, from each of the four addresses. On this coast the drive is part of the school.
What we will not measure
We will not rank on raw exam results, for the reason above. We will not rank on university destinations, which is the same problem with a longer lag. We will not rank on facilities, because a swimming pool is a capital decision rather than an educational one.
And we will not produce a single ordered list of best to worst, because the correct answer changes with the curriculum a family needs and the address they are choosing. What we will produce is a comparison across the criteria above, with the trade-offs named, so that a family can see which school is right for them rather than which is nominally first.
Why it has not run yet
Two reasons, and both are constraints we have chosen rather than obstacles we have hit.
The first is Article 04 of our charter: no institution is described from a prospectus. Ranking a school requires having been to it, spoken to staff who are not in admissions, and talked to parents whose children are currently there. That is a term of work, not a week of it.
The second is that the ground is about to move. Costeas-Geitonas has a binding long-term lease to open a primary and secondary school inside The Ellinikon in September 2027 — the first serious school physically on this coastline. A table published now would need reissuing within eighteen months, and the geography it describes is the part most likely to change.
What to argue with
The list above is the thing to disagree with, and now is the time. If staff turnover is the wrong proxy, we would rather be told before we spend a term collecting it. If there is a measure that matters more to a family in this position and is not on the list, it should be.
Write to the desk. A methodology published in advance and then quietly amended to fit the result is worse than no methodology, so this one will carry a version and a date, and any change to it will be published as a change rather than as a correction.
