Where the number comes from
52
Countries whose ratio moved by more than a point, and the test that it is not a modelling artifact
Change is measured within each country between 2000 and 2021, so nothing here depends on comparing one country to another.
| countries | |
|---|---|
| in the panel | 183 |
| moved > 1.0 | 52 |
| moved > 2.0 | 20 |
| median absolute move | 0.5119 |
the artifact test
If these swings were noise in WHO's estimation, they would be largest where the estimation does the most work. They are largest where it does the least.
| group | countries | median move |
|---|---|---|
| registration ≥90% complete | 72 | 0.7966 |
| all other countries | 74 + 32 + 5 | 0.4051 |
Mann-Whitney U, one-sided (counted > rest), p = 0.001794.
the countries that moved most
| country | 2000 | 2021 | change |
|---|---|---|---|
| Seychelles | 11.00 | 3.45 | −7.55 |
| Puerto Rico | 11.26 | 4.69 | −6.57 |
| Guatemala | 6.19 | 2.84 | −3.36 |
| Costa Rica | 7.73 | 4.60 | −3.14 |
| Barbados | 9.06 | 5.95 | −3.11 |
| Botswana | 5.09 | 2.09 | −3.00 |
| Jamaica | 5.33 | 7.91 | +2.58 |
| Maldives | 4.98 | 7.61 | +2.63 |
| Thailand | 2.82 | 5.51 | +2.68 |
| Cuba | 2.51 | 5.70 | +3.19 |
| Georgia | 5.02 | 9.08 | +4.06 |
| Kuwait | 1.41 | 8.39 | +6.99 |
WHO Global Health Observatory + WHO Mortality Database + UN WPPWHO GHO API, WHO Mortality Database (ICD-10), UN WPP 2024, US Census Household Pulse · diambil 2026-08-06