The data a university needs before accreditation season
An accreditation cycle does not ask for anything an institution should not already hold. It asks for it in a shape, over a period, with evidence attached. That is where the year disappears.
Framework requirements change between cycles and differ by council. Treat this as a guide to the shape of the work, and confirm current metrics against the framework that applies to you.
Every institution that has been through a cycle recognises the pattern. Six months before submission, someone is asked for programme-wise enrolment for the last five years. The registrar has it, but in three different section structures, because the department reorganised twice. Someone else is asked for the student-to-teacher ratio, and discovers that the definition used by the council counts sanctioned posts while the institution counts people who actually taught.
Nothing is missing. Everything needs reconstructing. The cost of a cycle is almost entirely the cost of reconstruction, and it is avoidable.
Three properties every metric needs
A number that satisfies an accreditation body is not just correct. It has three properties your day-to-day reporting probably lacks.
The formula is written down, including who is in the denominator. Two people must reach the same figure without conferring.
Tied to a census date, not to whenever the query ran. Enrolment on 30 September is a fact. Enrolment today is a moving target.
Every figure traces to a document a peer team can open: a minute, a register, an order, a signed list.
Eight metrics, and the definition that decides them
These recur across frameworks. In each case the argument is never about the count. It is about the denominator or the date.
| Metric | The question that decides it | Held where |
|---|---|---|
| {{ m.k }} | {{ m.q }} | {{ m.where }} |
Evidence is a filing problem, not a writing problem
The documents a peer team asks for already exist. A minute approving a syllabus revision, a circular announcing a workshop, an attendance list from a guest lecture, a purchase order for laboratory equipment. They exist in somebody’s inbox and in a physical file, and neither is reachable in the week the request arrives.
Attach the document to the record at the moment the thing happens, not at the moment it is asked for. A workshop with its attendance list attached on the day is evidence. The same workshop reconstructed in March from a photograph and a memory is a claim.
This single habit removes most of the panic. It also has a side effect worth naming: it makes the institution honest with itself, because an activity with no document attached is visible as such all year rather than at submission.
Five years is the real requirement
Most metrics are asked for as a trend, which means the hard part is not this year. It is that the structure changed underneath the data three years ago and nobody recorded the mapping.
{{ h.d }}
Record the mapping at the time of the change, in one line: this programme became those two, on this date, with this cohort split. A single line written in the month it happens saves a fortnight three years later.
A calendar that makes submission dull
Spread across the year, none of this is a project. It is four short exercises and one person who owns the definitions.
An institution that does this arrives at submission with an export and a folder. One that does not arrives with a committee and a deadline. The difference is not effort in the final six months. It is a definition written down five years earlier.
Bring your last submission and the metrics that hurt.
A consultant maps each one to where it would be captured continuously, with its evidence attached at source.