Correlation Is Not Causation: How Spurious Connections Fuel Junk Science

At its core, the correlation-causation fallacy arises because two variables can move together without one directly causing the other. A third, hidden factor may be at play. In the ice cream example, the lurking variable is warm weather: higher temperatures drive both ice cream purchases and increased swimming activity, which in turn leads to more drownings. No amount of statistical correlation between ice cream and drowning can prove a causal link. Yet in health, economics, and politics, such spurious connections are routinely presented as evidence, often to sell products, advance an agenda, or reinforce a preexisting belief.

Consider one of the most famous modern examples: the claim that vaccines cause autism. This idea originated from a small, flawed study in 1998, which reported a temporal correlation between the measles-mumps-rubella vaccine and the onset of autism symptoms in children. The timing seemed suspicious: many children first show signs of autism around the same age that they receive the vaccine. But correlation in time is not causation. Larger, well-controlled studies involving millions of children have since found no link. The real cause of the correlation? The simple fact that autism typically becomes noticeable in early childhood, exactly when routine vaccinations are scheduled. The coincidence of timelines, combined with parental anxiety, created a powerful illusion of causality that continues to fuel vaccine hesitancy decades after the original study was retracted.

Another common arena for spurious correlations is nutrition and lifestyle science. Headlines frequently claim that eating a particular food—say, chocolate, coffee, or red wine—lengthens or shortens life. These claims often rely on observational studies that find a statistical association. For example, people who drink moderate amounts of red wine tend to live longer than those who abstain. Does wine cause longevity? Or do wine drinkers tend to have higher incomes, better access to healthcare, and healthier overall diets? Studies that fail to control for socioeconomic status, exercise habits, and other confounders can produce misleading results. The same pattern appears with “superfoods”: blueberries are associated with better memory, but blueberry lovers also tend to exercise more and avoid smoking. The correlation may be real; the causation is not proven.

The problem intensifies when confirmation bias enters the picture. Once a person believes in a correlation—say, that 5G towers cause illness—they actively seek out any apparent co-occurrence that supports the idea. They may notice a neighbor who got sick after a tower was installed, while ignoring the thousands of people near towers who remained healthy. This selective sampling turns a weak or non-existent correlation into a personal certainty. The internet amplifies this by algorithmically feeding users more content that reinforces their beliefs, creating an echo chamber where spurious connections feel undeniable.

So how can a doubter navigate this minefield? The first step is to ask: Is there a plausible causal mechanism? If someone claims that a supplement cures a disease, ask how it works biologically. Without a mechanism, the correlation is merely a coincidence. Second, consider the possibility of a third variable. What else could explain the observed relationship? The classic example is the positive correlation between the number of storks observed in a region and the number of human births—both are high in rural areas with more farmland and families, not because storks deliver babies. Third, demand controlled studies. Randomized controlled trials, where participants are randomly assigned to a treatment or placebo, are the gold standard for establishing causation because they eliminate confounders. Observational studies can suggest hypotheses but cannot prove cause and effect.

Finally, embrace the uncertainty that comes with not knowing. Junk science thrives on absolutism: “This causes that, period.“ Real science thrives on nuance: “There is an association that needs further investigation, and here are the limitations.“ The doubter’s strength lies in resisting the urge to leap from correlation to causation, recognizing that just because two things seem connected does not mean one drives the other. By cultivating this skepticism—and by understanding the hidden variables that create spurious links—we inoculate ourselves against a vast swath of misinformation. The world is messy, complex, and rarely reducible to a simple X-causes-Y headline. And that complexity, once embraced, becomes the foundation of genuine critical thinking.