Professors teaching statistics like to tee up lessons on “correlation vs. causation” by giving examples of events that may appear statistically linked – but actually don’t have a cause-and-effect relationship. The most classic of these is the observation that ice cream consumption and murder rates spike at similar times, such as seen here:
It isn’t that ice cream leads to murder, or murder to ice cream.
Rather, the illusory correlation of these data points relies on a third variable: summertime. These warmer, more social months lead to both more ice cream eating and more violence.
Thank goodness ice cream is vindicated from homicidal suspicions – and the stats lesson moves on.
But since July 16 is National Ice Cream Day, we wanted to take a closer look at what this sweet treat may relate to – criminally or otherwise.
Our approach eliminates the potential that summer will artificially create the appearance of causality, by looking at ice cream data by year rather than by month.
Here’s what we found:
Indeed, annual ice cream consumption – as measured by the average pounds of ice cream consumed per person – has a strong relationship to murder rates in Florida between 1975 and 2015.
In other words, the more ice cream people consume in a given year, the higher you can expect the murder rate to be.
But there’s a twist that may once again vindicate ice cream.
The portion of homicides committed by firearms goes down in years where people consume more ice cream.
No – this doesn’t mean there are more murders by scooper in years where people have a higher appetite for ice cream. Again, this bizarre relationship can be explained by unrelated outside factors.
Ice cream consumption declined about 30% over the four decades between 1975 and 2015, from about 18.2 pounds to 13.1 pounds per person. In Florida, annual sales declined by about $500 million in candy, nut, and dairy store sales during that period.
In a similar way, Florida’s homicide rate declined 64% during that time, from about 14.4 to 5.2 murders per 100,000 residents. But murders by firearms jumped by a quarter from 1979, when they represented 59% of homicides, to 2016, when they accounted for 74% of homicides.
When you look at the whole picture, you can see how ice cream would appear to be related to criminality, when in truth it’s an innocent bystander.
Here’s some other (albeit completely spurious) good news about ice cream consumption: The more people eat ice cream, the more Florida Lottery winners there are each year!
We could probably debunk this lucky association … but why? After all, what could be so wrong about a world in which Moose Tracks and Powerball thrive in unison?