We build a daily game out of world datasets, which means we spend a lot of time looking at country rankings that are technically correct and completely misleading. The same handful of traps come up again and again.

Here they are, each with a real example you can check yourself.

1. A share of GDP is not an amount

The single most common trap. When an indicator is expressed as a percentage of GDP, the denominator moves — and in poor countries it moves a lot.

Worked example: Afghanistan spends 14.99% of GDP on health. Qatar spends 2.52%. Afghan life expectancy is 66.3 years; Qatar's is 82.5.

Afghanistan's high ratio is not generous healthcare. It is a small economy carrying a heavy disease burden. Across 193 countries, health spending share and life expectancy correlate at just 0.18 — essentially no relationship.

Before reading any "% of GDP" ranking as a measure of effort or quality, ask what happens to the ratio when a country gets richer.

→ The full breakdown: healthcare spending vs life expectancy

2. Check which way the index runs

Composite indices do not agree on direction, and ranking pages routinely sort them wrong.

Worked example: On the Reporters Without Borders press freedom index, lower is better — Norway scores 6.72, Eritrea 81.45. On Transparency International's corruption index, higher is better — Denmark 90, South Sudan 8.

Sort both "highest first" and you produce a leaderboard where Eritrea leads on press freedom. This is not hypothetical; it happens on published ranking pages regularly, including one on this site that we are fixing.

→ Press freedom and corruption, and why the correlation is an upper bound

3. "Renewable" may not mean what you think

Definitions inside an indicator matter more than the indicator's name.

Worked example: DR Congo gets 96.3% of its energy from renewables — the highest share on earth, on a GDP per capita of $649. Kuwait gets 0.1%.

The World Bank's renewable share includes traditional biomass: firewood, charcoal and dung. DR Congo's figure describes energy poverty, not an energy transition. Across 200 countries, renewable share correlates with income at −0.24 — negatively, because industrialising countries substitute fossil fuels for wood.

→ Why the world's "most renewable" countries are its poorest

4. Some ratios legitimately exceed 100%

If a number looks impossible, check the definition before assuming an error.

Worked example: Greece records 165.1% tertiary enrolment. This is a gross enrolment ratio — all enrolled students divided by the population of official university age. Mature students, repeat years and international students land in the numerator but not the denominator.

The consequence is practical: the difference between 97% and 142% is mostly noise, while the difference between 4% and 40% is real. Treat the top of such a scale as a band, not a ranking.

→ Fertility and education across 150 countries

5. A strong correlation is usually a proxy for something else

The tightest relationships in country data are rarely direct causes.

Worked example: Sanitation access and maternal mortality correlate at −0.80, stronger than income, doctors or health spending. Poor sanitation does cause some maternal deaths through infection — but it cannot carry a relationship that strong alone.

What sanitation coverage really measures at national scale is whether a state can deliver infrastructure to everyone. So does antenatal care. The plumbing and the midwives arrive together, because the same institutions deliver both.

→ Sanitation and maternal mortality

6. Correlations are carried by one end of the range

Check whether a relationship holds everywhere or only where things are worst.

Worked example: Physician density and infant mortality correlate at −0.73. But almost all of that is the gap between Niger's 0.04 doctors per 1,000 people and roughly 3 per 1,000. Above that, the line flattens: Belarus records 1.8 infant deaths per 1,000 with 4.72 physicians, better than Portugal's 2.7 with 5.85.

The correlation says a lot about the difference between having a health system and not having one, and almost nothing about the difference between a good one and a slightly better one.

→ Do more doctors mean fewer infant deaths?

The general rule

Before you quote a country ranking, ask three questions:

1. What is the denominator, and does it move?
2. Which direction is good, and is the table sorted that way?
3. Is this measuring the thing, or something that comes bundled with it?

Most misleading world statistics fail at least one of those, and almost none of them are wrong on the facts. They are just answering a different question from the one the headline implies.

If you want to practise reading choropleths quickly, that is more or less what our daily map game is — a world map shaded by a real dataset, and you work out which one. It builds the instinct faster than reading tables does.

All the data behind these examples is browsable in the atlas, with sources and years on every page.