Published on 2026-07-25
If two numbers rise together, does one cause the other?
Ice cream sales and drownings rise on the same days every year. One doesn't cause the other. Here's the habit of mind that keeps you from mixing them up.
Somewhere out there is a chart showing that the number of Nicolas Cage films released in a given year tracks, almost line for line, the number of people who drown in swimming pools. Tyler Vigen, a Harvard Law student, built an entire website called Spurious Correlations to collect graphs like this one: margarine consumption next to divorce rates, cheese eaten per person next to deaths by bedsheet tangling. Every pair rises and falls together with eerie precision. As Vigen himself puts it, the point isn't to make you distrust research, it's to make you look twice before you trust a chart.
Nobody thinks a new Nicolas Cage movie sends people to their pools to drown. The joke lands instantly. The trouble is that plenty of real, serious-sounding claims lean on the exact same kind of chart, minus the joke, and we don't always catch it.
The ice cream that doesn't drown anyone
The textbook version of this trap uses ice cream and drowning. Across a whole year, the months when people buy more ice cream are also the months when more people drown. Read only the chart, and you could build a scary headline out of it. Read the calendar, and the answer falls out on its own: both go up in summer, because it's hot, people eat more ice cream and people also swim more, which means more chances for something to go wrong in the water. Ice cream never touches the outcome. Heat drives both numbers separately.
That third, hidden driver is doing all the real work, and it never shows up on the original chart. Once you know to look for it, you start noticing how often two curves move together for a reason that has nothing to do with either one causing the other.
Four ways two lines can move together
Whenever two things rise and fall in step, there are only a handful of honest explanations, and it's worth running through them before believing the flashiest one.
- A actually causes B (the one everyone jumps to first)
- B actually causes A (the arrow can point the other way)
- A shared cause drives both, like summer heat driving both ice cream and swimming
- Plain coincidence, which becomes almost guaranteed once you compare enough random things
That last option matters more than it sounds like it should. Compare enough pairs of numbers, movies against weather, snack sales against sports scores, and pure chance will hand you convincing-looking matches on a regular basis. A website can be built entirely on that fact, which is exactly what Vigen did, deliberately, to make the point stick.
The word that should slow you down
Watch for 'linked to'. It's a phrase that lets a headline sound causal while staying technically honest, because a link is all a correlation ever promises. 'Sleep linked to memory', 'coffee linked to longer life', 'screen time linked to mood': every one of these might describe a real pattern in the data, and none of them tells you which way the arrow points, or whether something else entirely is pulling both strings.
The fix isn't to distrust every number you see. It's to ask one extra question before sharing: what's the mechanism? If eating more vegetables is genuinely good for the heart, there should be a plausible biological reason why, one that researchers can test directly rather than just spot in a chart. If nobody can explain how A would produce B, a strong correlation is a lead worth investigating, not a verdict worth repeating.
This is also why a single chart, however dramatic, is never the end of the story. Real researchers who suspect a genuine cause don't stop at 'these two things move together'. They try to rule out the shared cause, they compare groups that differ only in the one factor being tested, and they check whether the pattern still holds once the obvious third variable is accounted for. That extra work is invisible in a headline, but it's the entire difference between a correlation and a finding you can act on.
Nicolas Cage and pool drownings are a safe place to practise the habit, precisely because the absurdity is obvious and nobody's feelings depend on the answer. The real test comes later, with a claim that isn't obviously silly, where the two curves rise together and a serious voice is already telling you why. That's the moment to ask what's actually connecting them, rather than assuming the chart has already answered for you.