Evidence
Five kinds of evidence: a measurement, published theory, a model that reached people at scale, an institution that came out of it, and a working tool. Each is listed with what it demonstrates and what it does not.
The Fairness Index applied in full to one cohort.
| Domain | DFS | Reading |
|---|---|---|
| Governance and Leadership | 1.70 | Strong unfairness consensus |
| Justice and Rule of Law | 1.81 | Mostly unfair |
| Economy and Resources | 1.88 | Mostly unfair |
| Public Services and Welfare | 1.44 | Strong unfairness consensus |
| Accountability and Institutions | 1.85 | Mostly unfair |
The instrument produces differentiated, interpretable results on a real population rather than noise. All five domains fell below 2.00 — a broad consistent signal, not unfairness concentrated in one area.
This is not a measurement of Bangladesh. The cohort was 71 per cent aged 18–25, 62 per cent urban and 54 per cent undergraduate. It is a youth-educated-urban signal and nothing wider. No figure here should be cited as a national score.
Every question, every response distribution, every standard deviation and the full calculation chain are in the report of record, April 2026.
One result from the pilot carries further than the headline figure.
Eighty-five per cent answered completely unfair — the clearest consensus anywhere in the instrument. The subject maps directly to Article 17 of the Constitution of Bangladesh, which promises free and compulsory education. Article 17 sits in Part II, which is judicially unenforceable, and under Article 7B is not amendable by any means.
The provision exists. The agreement exists. No route connects them. That is what CARO calls Legal Injustice, and here it is measured rather than argued.
The theory the measurement rests on, in citable form.
| Work | Scale | Where |
|---|---|---|
| Fairness as Foundation | 380 pages | SSRN 6632960 |
| The Meta-Right to Structural Fairness | 41 pages, four authors across three universities | |
| Fairness as a Meta-Norm | 13 pages |
The framework is written down, complete and open to attack. Chapter Eleven is the constitutional blueprint; Chapter Thirteen specifies every formula the Index uses. The Meta-Right paper was written with researchers at the University of Dhaka, Shahjalal University of Science and Technology, and Bangladesh University of Professionals.
None of it has completed peer review. It is published and citable, not peer-certified.
What the six-stage sequence actually reached between 2024 and 2026.
| Stage | Reach |
|---|---|
| Civic awareness (MEP) | 450+ educational institutions, all eight divisions |
| Campus fellowship | 450+ applications from 143 campuses; 392 selected across 247 universities |
| Immersive training (VEP) | Four batches, 130+ selected, ages 16–35 |
| Field research (CMEP) | ~450 trained volunteers; 16 divisional leaders |
| The network built | 2,259 verified participants across 211 of 300 parliamentary constituencies |
People with no prior exposure to formal civic education can be trained to collect governance measurements to a standard, and the pipeline sustains itself — alumni returned at every stage and several went on to hold departmental roles.
Reach is not retention. Whether a local unit keeps measuring without the organisation that trained it is the open question, and the programmes ran while CARO was funding them.
The end of the sequence, and the only test that matters.
Founded under the Societies Registration Act 1860 by members subscribing a memorandum adopted from the formation kit. It elects its own officers and can remove anyone, including the people who founded it. CARO does not direct it, fund it as a controller or act for it.
Its registration is incomplete, its council is not fully constituted, and it has published no measurements. Whether a local unit is still taking readings in month thirty, with nobody checking, is the question this whole model exists to answer. It cannot be answered yet.
PlayerOne — built, running, and open to anyone.
A browser-based governance simulation that puts a participant inside a decision: a fixed budget, competing claims, a hearing to face, and every choice evaluated against structural fairness principles. It also collects Fairness Index responses in-simulation.
The measurement instrument can be deployed at distance, without field staff, and the training layer can run without an organisation present.
Cite the version and the cohort, never the organisation alone.
The cohort was 71 per cent aged 18–25, 62 per cent urban and 54 per cent undergraduate. Any citation carries that qualification.
None of them was paid a salary.
Students and volunteers across Bangladesh wrote the questions, reached participants in 211 constituencies, checked consent, cleaned the data and built the platform it was collected on. They are named in full on Built by volunteers.