Cool Correlations Explained

Correlations reveal relationships between variables, and are often described as a single number “r” falling between -1 and 1.

A negative correlation describes an inverse relationship, and a positive correlation describes a direct relationship.An r = 1 or -1 represents the strongest possible correlation, while an r = 0 is the weakest.

Correlations can therefore provide great insights into data, as well as spurious connections.

Consider the wonderful world of cheese.

(Delicious mozzarella cheese. Photo: Wikipedia Commons)

There is a correlation between per capita consumption of mozzarella cheese, and the number of doctorates awarded in civil engineering (r = 0.96). Even stranger is the correlation between per capita cheese consumption, and death by bed sheet entanglement.

One might assume that cheese has amazing causal powers, including the dualistic ability to kill and educate. This is a dangerous mistake.

First, a correlation does not prove causation. Eating a cheese pizza will neither result in a Ph.D., nor send your linens into a murderous rage. Second, the relationships between cheese, engineers and death are likely spurious. The connection may very well be by chance, though one can ask whether cheese and doctorates may be sensitive to an underlying trend.

Want to explore the world of spurious correlations in greater detail? Please visit the terrific website of Tyler Vigen.