Elite Careers StrategyWho Gets In?A research programme published by Elite Careers Strategy
London
Evidence dates 30 and 31 August 2026
Revised 9 September 2026
Research lead: Hassan Akram
For journalistshassan@accessecs.comInterview or comment within three working daysData checks same dayPress kitDownload reports
StandardOfficial primary sources only
No secondary reporting
No private client evidence

The register
Methodology

How this research is built, and what it cannot tell you

A register-based programme. Every published statement traces to one row in the Global Elite Selection Monitor, and every row traces to one official primary source, checked on a stated date.

The register

1:1one row records one verifiable selection signal
Aminimum source grade, official primary
107rows, each with source ID, URL, grade and review date
GradingGrade A - official primary
Confidence recorded 1 to 5
Review date recorded per source
Publication ruling per signal

Unit of analysis

One row records one verifiable selection signal: a requirement, stage, assessment, institutional statement or material change. Multiple signals may come from the same source. Each observation is narrowly worded to match its source rather than paraphrased upward.

Cell audit

Every positive cell in the Report 02 market matrix is published with the quoted passage that supports it, its source URL, the retrieval method and date, its verification status and its publication ruling: the cell audit (CSV). Absent cells were searched on the same terms as present ones, under the symmetrical review protocol, and each published sentence is traced to the observations it rests on in the claim dependency manifest (CSV).

Publication rulings

All 107 coded signals were re-checked against their cited source on 9 September 2026 with a full JavaScript render. Each carries a verification status and a publication ruling: published, published as context, or held. Six observations are held because their source now returns a bot-verification interstitial rather than its content; none is asserted in any report. The log is published as the publication-ruling log (CSV). For each of the 67 retrieved sources, the source evidence record (CSV) publishes the URL, the retrieval date, the character count, a SHA-256 of the retrieved text and each quoted passage with its offset, so the check can be repeated. The retrieved page text is not republished: the pages carry their own copyright.

Source standard

Sources are official employer or university pages. Report 01 draws on 48 official sources checked on 30 August 2026, held in a 50-entry source register. Report 02 draws on a purposive six-market sample checked on 31 August 2026, bringing the Monitor to 107 coded signals citing 69 official sources, held in a 71-entry register.

Coding

Report 01 coded signals to categories including live judgement, simulation, case or work simulation, multimodal or technical assessment, problem solving, gamified assessment or whole-person evaluation. Its 6/6 measure covers all observations coded as newly explicit. Report 02's market matrix is binary and conservative: a feature is marked present only where the cited source explicitly describes it.

Sample allocation

Eight per market at first publication on 31 August 2026. The United States and Hong Kong carry a ninth signal each, coded on 9 September 2026 while re-verifying the market matrix. Those two were found while looking for support for specific cells, so the sample is no longer equally allocated and the two additions are not a neutral addition to it.

Sampling

Report 01 is a purposive, cross-sectional sample weighted toward the UK, global careers pages, law and banking. Report 02 uses a purposive sample of eight signals per market at first publication, now UK 8, GCC 8, India 8, United States 9, Singapore 8, Hong Kong 9. Signal counts are close but no longer equal. Institution type, programme and applicant stage are not matched either, and the pathways compared are not equivalent.

BoundaryAI did not cause institutions to introduce live assessment. This claim is never strengthened, in any publication, briefing or interview.

Material limits

Sampling

These are purposive samples. Published process pages may omit informal practices and may describe different programmes that are not directly comparable. The dataset cannot establish prevalence, causal change over time, candidate outcomes or institutional intent beyond the stated process.

The geographic sample is too thin for league tables or country rankings. Percentages describe the coded sample only; they are not estimates for any wider market.

Report 02 compares published architecture. It does not rank markets, estimate selectivity, or infer fairness, intent or candidate outcomes. Live verification can protect integrity while also rewarding candidates with more time, coaching and familiarity; this research identifies that tension without measuring it.

Publication rule

Market-wide claims require triangulation across institutions and sectors. Historical trend claims require longitudinal evidence. Each report distinguishes what its sample shows from what remains a hypothesis. No proprietary index will be published until sample size and coding reliability support one.

AccountabilityResearch lead Hassan Akram, Founder and Principal Advisor, Elite Careers Strategy. Coding, publication rulings and corrections are his responsibility.

Independence and corrections

All published material is drawn from public institutional sources. No private client evidence is used in any report. Nothing in this programme implies affiliation with, or endorsement by, any institution referenced. Where an institution is named, the accompanying statement is confined to what that institution itself publishes.

Institutions change their published processes. If a signal attributed to your organisation no longer reflects your current published process, write to hassan@accessecs.com with the current source page. Corrections are made against the source and the review date is updated in the Monitor.

Revision 9 September 2026

Reviewed and approved by the research lead on 9 September 2026. Rulings marked proposed are not yet approved and are identified as such in the log.

Area
Change
Reason
Status
Exhibit 3
Rebuilt again
Every cell rechecked against the observation it cites, not only against its source. Four cells cited a source the register assigns elsewhere, two rested on observations that cannot support their row, and the GCC prior-record cell had no supporting observation at all and is withdrawn. Verified present is 26 after a symmetrical review of all 36 cells, and two rows rather than three are verified in all six markets.
Applied 9 Sep 2026
SIG-106, SIG-107
Coded
Two observations were coded on 9 September because the source states the thing and no existing row recorded it: Freshfields on academic record in Hong Kong, and Harvard on Admissions Committee review. Both were found while looking for support for specific cells, so they are not a neutral addition to the sample.
Approved 9 Sep 2026
Counts
Restated
The active register is 107 signals, not 105, and Report 02 draws on 50, not 48. The allocation is no longer equal at eight per market: the United States and Hong Kong carry nine each. No observation was superseded to preserve the old figures, because retiring a true observation to hold a count would be selection toward the result.
Applied 9 Sep 2026
Rulings
Approved
Six held observations approved as held and excluded. Twenty-one context-only observations approved as context only; none may underpin an affirmative claim. The remaining rulings stand as proposed. Every ruling, with its basis, retrieval method and approval status, is published in the publication-ruling log.
Approved 9 Sep 2026
Evidence
Published
A source evidence record publishes the URL, retrieval date and method, character count, SHA-256 and quoted passages with offsets for every retrieved source, so any reader can repeat the check. The retrieved page text is not republished.
Applied 9 Sep 2026

Revision 8 September 2026

Area
Change
Reason
Status
Exhibit 3
Rebuilt
Each cell re-checked against its official source page and against the registered observation it cites. Verified present cells fell from 32 to 25.
Applied
Exhibit 3
Held
Every cell was rebuilt against a verbatim passage from the source rather than against the source alone. Verified present fell to 23 at this revision; the symmetrical review that raised it to 26 was carried out on 9 September and is recorded in that revision.
Applied
Exhibit 3
Relabelled
"Timed test" became "Test or structured assessment". Timing was evidenced in one market only, so it is no longer asserted across the row.
Applied
Headline
Withdrawn
"Four of the six evidence types appear everywhere" is not supported by the rebuilt matrix. Two types are verified in all six markets: test or structured assessment, and interview. Prior record or academic threshold is verified in five.
Applied
Chronology
Corrected
"50 of 57 predate generative AI" removed. The change-type field records what a page said at the review date and carries no date of introduction.
Applied
46 per cent
Withdrawn
Both the percentage and the underlying count are withdrawn. Verified institutional examples are published instead, pending a published inclusion rule and row-level review.
Applied
Framing
Corrected
"Matched cohort" became "purposive six-market sample". Institution type and pathway are not matched, and since 9 September 2026 the signal count is no longer equal either.
Applied

Evidence dates are unchanged: Report 01 checked 30 August 2026, Report 02 checked 31 August 2026. The 8 September 2026 re-verification checked the coded observations against their official source pages; it did not re-date the evidence.

How this revision was produced: the source pages were retrieved and compared against the register with machine assistance, and every resulting change was reviewed before publication. Coding decisions, publication rulings and corrections remain the responsibility of the research lead.

Versions

Publication
Version
Change
Date
Report 01
Revised
Source count restated as 48 cited sources held in a 50-entry register.
31 Aug 2026
Report 01
First published
57 signals, checked 30 August 2026.
30 Aug 2026
Report 02
First published
Published as 50 signals across six markets, checked 31 August 2026.
31 Aug 2026
Register
Published
Full signal log and source register published at /who-gets-in/register/.
8 Sep 2026

Corrections issued to date are listed above. Where a correction changes a published figure, the superseded figure is named rather than removed. Signals are re-checked on the review date carried in the register.

Suggested citation

Akram, H. (2026). Methodology and limits. Who Gets In?, Elite Careers Strategy, Global Elite Selection Monitor. www.accessecs.com/who-gets-in/methodology/