People killed by lightning

As Americans have eaten more ice cream per capita, more of them have been killed by lightning, the most famous confounding variable example in all of statistics and the reason every introductory course begins with a warning about causation. The coefficient is 0.876 across seventeen years, and the explanation has been the same since approximately forever: both things happen more when it is hot outside. The sun melts the ice cream, heats the atmosphere, generates the storms, and the data generates the correlation. Summer is the confound that launched a thousand lectures.

Both bent in 2020 from the same cause. Ice cream consumption jumped as households stocked freezers for lockdown comfort, while lightning deaths fell because a pandemic-locked population was spending dramatically less time outdoors in thunderstorm weather. The shared variable is how many Americans were inside on any given summer afternoon.

Seventeen years of ice cream and lightning is the correlation that every statistics student learns first and forgets last. The confounding variable is summer itself—the season that produces both frozen treats and thunderstorms—and the lesson is that two things moving together need not have any awareness of each other's existence. The cone melts, the lightning strikes, and the temperature is responsible for both. Always has been.