Evidence people can use.
Select a country. Inspect its evidence. Build an assessment and see how the original equations respond.
Human dignity is inherent. These observations describe conditions and institutions; they do not measure anyone’s worth.
Statistical observation years 2020–2024 · Snapshot assembled 25 October 2026 · English edition
Loading the assessment and its versioned evidence records…
Make the evidence and the calculation visible.
Select a country and period. Eligible reference responses load automatically; add or revise responses in your own assessment. The dashboard updates as you work.
Start with a working example or your own evidence.
See all three formula families calculate, then return to your own assessment. The example uses hypothetical responses and keeps your work separate.
C01 is the displayed educational proposal: factor codes 1–5, exponent codes 1–2, Absent = 1. Starting with C01 explicitly selects this rule for your assessment.
Review the numerical rule · proposed C01
The original numeric mapping, weights and aggregation are unselected in the saved specification. C01 is a concrete proposal for KJA’s review. You can explicitly use it for exploratory scenario arithmetic now; this does not approve a KJA country measurement scale.
| Response | F, W, AF_ARMED, CE, SF_SECURITY, SF_SUSTAINABLE | I exponent code |
|---|---|---|
| Absent | 1 | 1 |
| Weak | 2 | 1.25 |
| Mixed | 3 | 1.5 |
| Substantial | 4 | 1.75 |
| Strong | 5 | 2 |
| Unknown / unanswered / N/A | Unavailable; no number, no denominator removal | |
For each equation input, average the 10 codes in each original domain, then average the six domain means with weight 1/6 each. All statements are coded by realization of the stated condition; no silent reverse coding. Adaptive Force, Human Dignity and Human Flourishing retain categorical profiles and do not enter these equations.
Below 50 substantive responses, or below 8 in any input domain, a component is incomplete. With 50–59 and domain coverage met, show a completion range over unresolved codes, without a point estimate. With all 60, show the component point value. A range is a mathematical completion bound, not a confidence interval.
Positive-floor choice: Absent is 1, not zero. Thus factors are 1–5, I is 1–2 and raw outputs are 1–625. This keeps the exponent direction nondecreasing within this proposal but means an absent condition does not annihilate the output. Equal spacing, this floor and equal weights are modelling choices, not empirical findings.
Results retain full JavaScript numeric precision internally; the dashboard displays up to eight significant digits. There is no percentage, probability, ranking, forecast or “good/bad” cutoff. Compensation in a mean cannot cancel a rights violation. Human dignity is inherent.
Three families: V = (F × W)I; SCR = (AF_ARMED × CE)I; SV = (SF × CE)I, with Security and Sustainable profiles kept distinct. Adaptive Force is never substituted for Armed Force.
Proposal KJA-MEE-EDU-C01-P1 · based on saved C01 research candidate · KJA approval not recorded. Independent empirical validation is deferred for educational operation and remains a separate research status.
“Live” means recalculation from this assessment’s stored inputs. It does not mean continuous monitoring of countries. Changing a scenario input shows a mathematical effect, not evidence of a real intervention’s effect.
Completion across ten banks
Substantive responses out of 60. Select a bank to inspect the contributing facts. Counts describe completion, not evidence quality.
Where responses come from
All 600 facts in the active context. Each fact appears once. Select a category to inspect its facts.
Component values on the proposed scales
Dots show complete component values; lines show conditional completion ranges. Select a component to inspect its facts. Factor codes and the I exponent use separate labelled axes. Missing components are not plotted.
Component values, domain calculations and reference comparison
Compare reference and scenario only within the same equation, country, dates, profile and proposed rule. Different equations have different meanings. No cross-country ranking is produced.
No compatible multi-period reference series is available in this release; no MEE historical trend is invented. Original source-statistic trends remain in Explore facts.
The original facts, one decision at a time.
Category definitions and daily-life meaning
Accessible fact matrix and domain counts
Legend: S Strong · Su Substantial · M Mixed · W Weak · A Absent · ? Unknown · — Unanswered · N/A Not applicable. Select a fact to open its response and evidence. Colours are supplementary.
Build an explicitly hypothetical scenario more quickly
Choose a bank and a hypothetical category. This fills only its unanswered facts, preserves reference responses and existing answers, and marks every filled entry as hypothetical. It is not evidence about a country.
Save, resume and reproduce
Responses stay in this page’s memory until you save. Export a complete record to keep it; imported records are recalculated against their exact model and reference version. Saving in this tab uses session storage on this browser and is optional. No assessment response is submitted to KJA.
A shared browser can expose a saved tab record to its next user. Export it and use Forget tab save when finished. Exports contain your notes and evidence links. The surrounding website may use its own cookies.
Reference eligibility, evidence coverage and release status
Look at a place, a year and the evidence.
Values retain their original units. A national statistic can reveal a condition without establishing every legal protection, lived experience or subgroup disparity.
Missing means not reported in this captured snapshot, never zero. Estimates may be revised. The display rounds to six significant digits; expand a source card or download the view to inspect the captured precision. Country and territory labels follow the provider.
What does KJA MEE ask?
Search all ten original banks. A linked statistic is partial evidence for a question, not an automatic rating. No condition-to-number mapping is applied here.
Build a traceable evidence profile.
Select a place, statistical year and component. Inspect all 60 unchanged questions alongside the linked observations and the evidence still required. This is an evidence inventory for review; no whole-question condition or numerical MEE input has been accepted.
This inventory covers evidence embedded in this workbench only. An empty entry means no matched observation here; it is not a finding that the condition is absent. Documentary context has not been qualified for the selected statistical year. Evidence tests are author research proposals. The original bank rubric remains intact; independent empirical validation is separately preserved and deferred as a prerequisite for educational calculator operation. Downloads stay on your device and are not automatically submitted or accepted.
Read the proposition beside its limits.
Inspect 111 bounded observations: 85 retained findings plus 26 source findings on South African justice, labour inspection, detention, cultural and religious complaints, and environmental enforcement. Each retains its scope and limitations. Statistical survey years, financial years, event dates and publication dates are distinct.
Try the original equations.
Why the scale of an input matters
For positive A × B, changing I has the direction of log(A × B). Increasing I lowers the raw result below a product of one, leaves it unchanged at one and raises it above one. This mathematical property does not establish how better institutional conditions should be encoded.
The accepted research decisions require defensible scales, mappings, weights, uncertainty and empirical testing before national numerical MEE ratings. This demonstrator supplies no such acceptance.
Use evidence with care and purpose.
- Begin with a concrete question. Identify the right, institution or condition. Select a place and observation date.
- Inspect what the source measured. Check definitions, units, scope, modelled estimates and the original producer. Open the official source.
- Keep distinct claims distinct. A legal protection, its implementation, an observed outcome and an effective remedy require their own evidence.
- Look for gaps and unequal access. National values can hide exclusion. Missing records are not evidence that a problem is absent.
- Explain the calculation. State the inputs and assumptions. A correctly evaluated formula can still have unvalidated empirical interpretation.
- Use knowledge to improve conditions. Ask what lawful, rights-respecting action could address the documented issue and what evidence would show improvement.
What is available now
Country-year statistics, source metadata and attribution; the original 600 questions; 111 bounded source observations; comparison and export; arithmetic with explicit hypothetical inputs; and an evidence-linked user assessment dashboard with opted-in C01 proposal calculations.
What has a separate release requirement
Validated national MEE ratings need qualified complete profiles, accepted measurement rules, actual independent findings, frozen predictions and unseen empirical validation. These requirements have not been marked complete.
Sources, reuse and attribution
Statistical source: World Bank, World Development Indicators, with the original producers credited in each source card. Public numerical series have per-indicator metadata identifying CC BY 4.0. Read the World Bank additional terms and attribution requirements. The World Bank and original producers do not endorse KJA MEE. Statistical data retain their stated open-data licences; general site copyright notices do not restrict those third-party reuse rights.
The air-quality series is available using the official World Bank indicator page’s source and reuse statement, reconciled on 25 October 2026. Its captured values are unchanged; 2024 observations are unavailable in this snapshot. The older, mismatched metadata PDF is preserved in the research archive. Source observations are AI-assisted research summaries; the source and limits remain visible. New South African findings concern 2023/24 or 2024 and are not current 2026 assessments. Stats SA survey values use their native definitions; they are not World Bank series or MEE scores. The AU report and Auditor-General finding are published external evidence about elections and departmental reporting. Neither is an independent review of the KJA MEE numerical model. Two Stats SA PDFs were inspected online but could not be archived; their checksum fields remain unavailable.
Downloads contain the selected statistical records and attribution, a complete 60-question evidence inventory, or your hypothetical calculation. The inventory identifies partial support and outstanding checks without assigning ratings. This workbench has no learner account and sends no entered assessment values to a server. Assess & calculate offers explicit JSON save/resume and optional tab-only session storage. The surrounding website may have its own cookies and privacy policy.
Universal Declaration of Human Rights · official English · KJA dignity and evidence guide · Full formula and fact-bank reference