A university directory may show programmes operating today, including several with long histories. It is tempting to use that list to say how reliably programmes remain available. But a present-day inventory can omit programmes that closed, merged or changed identity. Selecting only the programmes still visible changes the denominator before any percentage is calculated.

The Office for National Statistics defines its three-year business survival indicator with reference to businesses born in a starting year. That is an example of a starting-cohort denominator, not evidence about universities. This article develops an independent fictional programme-ledger method. No business survival rate is transferred to colleges, and none of the programme counts below describe an actual institution.

Reconstruct the starting cohort

Imagine ten programmes first opened in year 1. By the end of year 3, six are still operating under unchanged identities, two have closed, one has merged into another programme and one has an unresolved record. Meanwhile, four new programmes have opened after year 1.

Group Count Treatment in the original cohort
Unchanged and operating 6 Confirmed continuing
Closed 2 Confirmed not continuing
Merged 1 Apply an explicit continuity definition
Unresolved 1 Retain as unknown
Opened after year 1 4 Not members of the starting cohort

A current directory might contain ten entries: the six continuing originals plus four later entrants. Dividing ten current entries by ten original entries produces 100%, but it does not measure survival. The numerator contains different programmes. It measures neither an unchanged cohort nor a verified outcome for every original member.

Define what continuing means

Under a strict definition requiring the same identity, six of ten starting programmes are confirmed continuing: 60%. If documentary evidence establishes that the merged programme counts as continuing under a broader declared definition, seven of ten are confirmed continuing: 70%. The unresolved record remains unresolved in both calculations.

Those figures are classification results for the invented ledger. They do not estimate a student's chance that their chosen course will close. Different entry years, institutions, programme types and continuity rules can produce incomparable results. A renamed course also should not be marked closed solely because a title search fails.

For a transparent uncertainty statement, distinguish verified outcomes from possible ones. With the merger accepted, confirmed continuity is seven; if the unresolved entry later proves continuous, it becomes eight. The corresponding arithmetic bounds are 70% to 80%, conditional on that merger definition and the known records. These are not sampling confidence limits.

Preserve identity evidence

Use stable identifiers where available, along with institution, award, campus, intake and official effective dates. Save the evidence connecting an old programme to a replacement rather than assuming similar names establish continuity. A merger might preserve teaching for enrolled students while ending new admissions to the old award. Which outcome matters depends on the question.

Ask separately about teach-out arrangements for current students and availability for a future intake. A directory entry can remain visible for a cohort finishing its studies even when new applications are closed. Conversely, an announced programme may appear before any student has enrolled. Neither status should be compressed into a single unexplained active flag.

Use the ledger for the decision it can support

For a real shortlist, request current intake information and any published continuation or closure arrangements from the institution. A historical cohort analysis can identify questions to investigate, but it cannot guarantee future provision. Keep the observation date and the meaning of every status in the report.

The practical output is a matched starting-cohort table with explicit identity decisions, later entrants excluded, and unknown outcomes retained. A present-day programme count remains useful as an inventory. It becomes evidence about continuity only after the original members and their subsequent outcomes have been traced under a consistent definition.