The published standard behind every mark.
A single, weighted, public document. Every accreditation and every scorecard the Institute issues is produced against it — no sponsored placements, no pay-to-rank.
The Standard exists to settle one question: would this decision survive scrutiny? When a board, a candidate, or a regulator asks how a hiring, promotion or succession decision was made, the answer should be a document — not an opinion. The GTI Talent Standard is that document: an open, weighted rubric that defines what good looks like and how it is measured, applied identically to every subject we assess.
It is deliberately narrow. The Standard does not rank brands, reward marketing budgets, or reflect popularity. It scores evidence against fixed anchors, so that two independent reviewers reach the same place — and so that any score can be traced back to the criterion, the weight and the evidence that produced it.
The weighting matters as much as the criteria. Validity carries the most weight because a tool that does not predict is not defensible, whatever else it does well. Fairness and reliability follow, then transparency, cost and faking resistance. The full weighting is set out below.
Six criteria. One hundred points.
Each subject is scored 0–100 against published anchors on every criterion, multiplied by its weight, and summed. The same anchors are applied to every tool in a category.
| Criterion | What it checks | Weight |
|---|---|---|
| Predictive & construct validity | Does the tool measure what it claims, and does it predict on-the-job outcomes? Evidence from peer-reviewed and technical manuals. | 30 |
| Reliability | Test–retest stability and internal consistency. Would the same person get the same result next month? | 20 |
| Fairness & adverse impact | Group-difference and adverse-impact data; legal defensibility for selection use (EEOC / equality law). | 20 |
| Transparency & evidence | Is the technical documentation, norm data and scoring logic openly published and independently auditable? | 15 |
| Total cost & usability | All-in cost including certification, per-report and platform fees — weighed against report quality and ease of use. | 10 |
| Faking resistance | How well the instrument resists impression management and deliberate distortion. | 5 |
Tiers are awarded on the total weighted score, with minimum thresholds on fairness and validity that cannot be averaged away — a tool cannot buy back a fairness failure with a strong cost score.
We score the evidence, not the brochure.
A score is only as good as what it is built on. Every verdict draws on the same four evidence sources, weighed in the same order.
- Published technical manuals and norm data, read against the same five figures every time: criterion validity, test–retest reliability, internal consistency, the norm sample, and adverse-impact ratios.
- Independent, peer-reviewed research on the instrument and its underlying constructs — weighted above vendor-commissioned white papers.
- Group-difference and adverse-impact evidence, assessed for legal defensibility in selection use under EEOC guidance and equality law.
- The publisher's own submission, scoring logic and live platform — examined directly, not taken on trust, with follow-up questions kept on record.
No vendor pays for a verdict.
Every assessment is formed by the external panel against the published rubric. No sponsorship, no pay-to-rank, no pre-publication review — the score is the score, and the weights above let anyone audit how it was reached.
A living standard, with a right of reply.
The science moves; so does the rubric. The Standard carries a version and a date. When weights, anchors or thresholds change, the version increments and the change is recorded in the open. A scorecard always names the version it was scored under. The current edition is The GTI Talent Standard v1 · 2026.
Appeals
A publisher or employer who believes a verdict is wrong can submit new or corrected evidence. The appeal is reviewed by panel members who did not produce the original score; the outcome — upheld or revised — is recorded against the scorecard. Marks are year-dated and revocable: a score reflects the evidence at the time of review, and nothing more.
Honest framing: GTI is new. We are recognising our inaugural cohort in 2026 and publish no results we have not yet earned. Where the methodology is still maturing, we say so.