Evidence before conclusions.
World Equality Index investigates difficult questions without requiring a preferred answer. This charter is the standard we commit to, and the standard readers can hold us to.
Equality is not one variable.
Equal rights, equal opportunity and equal outcomes describe different things. We keep those questions separate before examining the relationships between them.
- Rights
- Legal protection, civil liberties, equal standing and due process.
- Opportunity
- Access to education, capital, institutions and the possibility of advancement.
- Outcomes
- Observed differences in income, health, education and other results.
- Contribution & burden
- Paid and unpaid work, risk, taxes, benefits and responsibilities.
- Resilience & dependency
- Essential functions, dependencies, replacement and response to disruption.
We do not treat an outcome gap as sufficient evidence of discrimination, biological cause or merit. Each explanation needs its own evidence and research design.
Say what kind of claim it is.
A statistic, an explanation and a value judgment carry different evidentiary obligations. Every research document and claim identifies its primary claim type.
| Type | Question | Example |
|---|---|---|
| Descriptive | What was observed? | The measured share of women in an occupation. |
| Associational | What varies together? | An association between income and an outcome. |
| Causal | What changes what? | The effect of a policy under an identified research design. |
| Predictive | What can be forecast? | A validated model’s estimate of a future outcome. |
| Counterfactual | What if conditions changed? | A workforce absence scenario with explicit assumptions. |
| Normative | What should we value? | The choice to give equal rights a particular index weight. |
What we commit to.
Let the hypothesis fail.
Define what could support, weaken or falsify a claim. Report null findings and inconvenient results alongside favorable ones.
Keep the source attached.
Identify the source, release period, population, definition and material caveats behind a public number. Preserve reproducible inputs and transformations.
Keep uncertainty visible.
Report missing data, selection limits and uncertainty. An unexplained residual stays unexplained. A correlation does not become a causal effect through stronger wording.
Respect the level of inference.
Group averages cannot determine an individual’s ability, character or worth. Racial and ethnic categories depend on context and measurement, and are not automatically proxies for genetic ancestry.
Separate analysis from values.
Show the choices behind normalization, weights and aggregation. Test sensitivity. State value judgments explicitly instead of embedding them in an apparently objective score.
Protect the result from its sponsor.
Disclose material funding and conflicts. A sponsor must not have authority to approve, suppress or rewrite the conclusion.
A visible path to publication.
Frame the question
Specify the construct, scope, hypothesis and possible falsifier.
Register the sources
Document provenance, definitions, period and limits before analysis.
Write and version
Produce a manuscript with explicit findings and limitations. Save revisions.
Challenge the work
Seek both supporting and critical reviews of the same revision.
Release transparently
Publish a version with its sources, review provenance and release note.
Protocols describe planned work. Working papers are provisional analyses. These labels do not imply independent peer review. The studio can record reviews supplied by the editor; recording a name does not independently verify a reviewer’s identity.
A correction is part of the record.
We preserve the public release while a new draft is edited. A substantive change creates a new manuscript revision and requires review again. Release notes explain the public change.
This platform begins with U.S. workforce observations and research protocols. It does not yet estimate civilizational survival time or publish a measured global composite equality ranking.