Media Literacy · August 29, 2026
Correlation Is Not Causation: Why Anecdotes Aren’t Evidence
SpongeBob, autism, ice-cream sales, and shark attacks make one point: a pattern can be real while the causal story attached to it is completely wrong.
A story is not a control group
The video opens with a deliberately absurd argument. A friend who was not allowed to watch SpongeBob SquarePants was later diagnosed with autism. Laynie and her sister watched the show and were not diagnosed. If personal stories established causation, the obvious conclusion would be that watching SpongeBob prevents autism.
It sounds ridiculous because the reasoning is easy to see when the conclusion is unfamiliar. But the same structure appears constantly in health claims, political arguments, conspiracy theories, product testimonials, and social-media debates: one thing happened, another thing happened, and a confident speaker declares that the first caused the second.
Two lines can rise together without causing each other
The second example compares shark attacks with ice-cream sales. Both rise during the summer. That does not mean “big ice cream” is causing shark attacks, no matter how neatly the lines on a chart appear to move together.
The missing piece is a third variable: hot weather. More people buy ice cream because it is hot. More people swim because it is hot. More swimmers create more opportunities for shark encounters. The relationship between ice-cream sales and shark attacks is a correlation explained by another factor, not evidence that one causes the other.
What a causal claim actually needs
“I know somebody who…” can be a reason to ask a question, but it is not a scientific study. A useful investigation asks whether the result holds across a larger and appropriately selected group, whether comparison groups differ in important ways, whether the supposed cause came before the effect, whether other variables could explain the pattern, and whether repeated studies point in the same direction.
That does not mean personal experience is worthless. Experience can reveal a problem that deserves study. It can show how an outcome feels or how a system affects an individual. What it cannot do by itself is establish a general causal rule for everyone else.
A quick test for viral claims
Before building a worldview around a cousin’s diagnosis, a neighbor’s story, or a dramatic graph, slow the claim down. What exactly is being compared? Is the speaker showing causation or only timing? Is the example representative? Is there a plausible third variable? And does the actual research support the conclusion?
The video’s joke works because “more ice cream, more sharks” sounds like data until the mechanism is examined. Numbers do not lie, but people can give numbers a causal story the numbers never proved.
Edited transcript
The transcript is lightly edited for punctuation and clarity.
My friend wasn’t allowed to watch SpongeBob growing up, and then he got diagnosed with autism. But I watched SpongeBob, and my sister watched SpongeBob, and neither of us got diagnosed with autism. So clearly watching SpongeBob prevents autism.
Actually, you know what else I just noticed? Shark attacks spike in July. And you know what else spikes in July? Ice-cream sales. So obviously big ice cream is causing the shark attacks. Think about it: more ice cream, more sharks. The numbers don’t lie.
Except that is not how research works. An anecdote is not a control group. Two things happening at the same time does not mean one caused the other, and “I know somebody who…” is not a scientific study.
Sometimes there is a third variable, like summer. More people buy ice cream because it is hot. More people swim because it is hot. More swimmers mean more opportunities for shark encounters.
Before you build an entire worldview around something that happened to your cousin’s neighbor’s hairdresser, maybe check the actual research.