A headline shouts that a condition “affects millions.” True at national scale, useless at personal scale without context. Numbers impress because brains compress them into feeling, not because they automatically tell you what to do Monday.
Journalists work under space limits; editors love integers with commas. Researchers speak in rates, confidence intervals, and populations defined narrowly. The translation step is where zeros get lost or added in your head.
Checks before you react
| Phrase in article | Ask this |
|---|---|
| “Up 200 percent” | From what baseline? Small baselines explode percentages |
| “Millions exposed” | Over what time span and geography? |
| “Doubles your risk” | Absolute risk from what to what? |
| “Nation spends billions” | Per year or decade? Per person? |
| “Record high” | Records since when? Better instruments change records |
Per capita versus totals diverge on populous countries. A policy affecting one in ten thousand nationally still touches thousands of families—both statements can be true while sounding contradictory.
Lifetime risk and annual risk differ. Mixing them makes cancer or accident stats feel either trivial or apocalyptic.
Round numbers in PR are sometimes estimates with wide error. Precision to the second decimal in a press release is a flag, not proof of rigor.
Compare like with like. Vaccine trial thousands versus census millions are different universes. Good reporting says which universe.
Log scales on charts flatten or steepen trends depending on choice. Axis labels are part of the argument.
When a number matters for your decision, chase the primary source table. When it does not, muting is fine. Not every statistic deserves rent in your anxiety budget.
