Seventeen Country Indicators, But Far Fewer Independent Facts
Under-15 share and fertility rate correlate at 0.95 across 194 countries. Several of our own rankings are close to the same list in a different order.
Across all 194 countries, under-15 population share and total fertility rate correlate at 0.95, and six of the seventeen indicators on this site correlate with each other above 0.75 — so the rankings carry far fewer independent facts than their count suggests.
This site publishes seventeen country indicators. Correlate them against each other across all 194 countries and a lot of them turn out to be measuring the same thing.
The strongest relationships
| Pair | Countries | r |
|---|---|---|
| Under-15 share × fertility rate | 194 | 0.95 |
| Median age × fertility rate | 194 | −0.87 |
| Life expectancy × under-15 share | 194 | −0.87 |
| Life expectancy × fertility rate | 194 | −0.83 |
| Internet use × life expectancy | 193 | 0.77 |
| Secondary completion × under-15 share | 189 | −0.76 |
| Secondary completion × fertility rate | 189 | −0.75 |
| Internet use × fertility rate | 193 | −0.75 |
Under-15 share and fertility correlate at 0.95. That is close to measuring the same quantity twice. The under-15 ranking and the fertility ranking are, in practice, one list presented two ways.
That is not a flaw in the data — it is a fact about how populations work. A country with high fertility has a lot of children in it. But it does mean a reader comparing "two independent measures" is comparing one.
The single factor underneath
Most of that table is one thing wearing eight hats. Countries with high fertility have young populations, shorter life expectancy, lower school completion and less internet access — and the reason is not that any of those causes the others directly.
Almost all of it tracks income and development stage. Internet use correlating with life expectancy at 0.77 is not a claim that connectivity extends life; it is two measurements of how rich a country is.
The two that do not correlate, and why they matter more
The weak relationships are the interesting ones, because both are places where policy assumes a link that the data barely supports.
Compulsory years of schooling × secondary completion: r = 0.44. Mandating more years is only weakly associated with more children finishing. The mandate ends anywhere from age 10 to age 22, and a law without schools, teachers or affordable access is a statement of intent rather than a delivery mechanism.
School starting age × secondary completion: r = −0.14. Essentially nothing. Starting children at four rather than seven does not predict whether they finish. That is consistent with the outcome evidence on early school starts, and it is worth knowing before treating an early start as a head start.
What to do with a ranking
Three practical rules follow.
Do not stack correlated indicators as if they were evidence. Citing under-15 share, fertility and median age together is one fact repeated three times with three sources attached.
Prefer the indicator closest to the thing you actually care about. If you mean "are there a lot of children here", use the under-15 share. If you mean "how fast is that changing", use fertility. They are not interchangeable at the margin even though they correlate at 0.95.
Treat the weak correlations as the informative ones. Where two things that ought to be linked are not, something is happening — and that is usually where the real question is.
See all seventeen at the data hub, the regional aggregates at /regions, or download the whole table at /countries.csv.