Definition
A lonely number is Rosling's name for a single statistic presented with no comparator, no denominator, and no trend — "Last year, X people died of Y." A lonely number can sound enormous or trivial without being either, because there is nothing in the sentence to anchor the scale.
The corrective is to refuse to react until at least one comparator arrives — a per-capita rate, a baseline year, an alternative cause, or a population in the same bucket.
Why it matters
How it works
Three classes of comparator dissolve most lonely numbers. The first is a denominator: deaths per population, dollars per worker, cases per 100,000. The second is a baseline: how does this year compare with the long-run average? The third is a peer set: how does this country, age group, or sector compare to its neighbours on the same metric?
When you see a lonely number in a headline, write it down and then ask which of those three is missing. The answer is usually all three. Find one comparator and the number stops being lonely — it becomes a data point, with all the constraints and qualifications that implies.
Rosling's own routine when reading news was to mentally append "out of how many, compared to last year, and compared to which alternative?" to every figure. He claimed this single habit reframed about a third of the news he encountered.