Correlation is a necessary component of causation. You observe two things and try to figure out a relationship. Your conclusions are limited because you can’t prove anything with a correlation but it’s a good place to start.
Correlations don’t prove causation. A can cause B but it is just as likely that B causes A. And it is equally likely that C caused both A and B.
Franz Gall gives us two examples. Gall was a comparative anatomist. He used the skull size of animals as his independent variable (input) and musket balls as his dependent variable (outcome measure). He took skulls of different sizes and poured the musket balls (lead shot) into each.
He reasoned that the more lead shot a skull could hold, the larger its skull was and the larger its brain. And larger the brain, the smarter the animal would be. Gall was mostly right.
There is a minimum size required and a maximum. At some point it is not the size of the brain but how it is constructed that matters. But it wasn’t a bad guess. There is a correlation there, at least in the middle region.
Gall’s next correlational study was between the bumps on the head and the mental abilities of people. He thought of the brain as a muscle that gets bigger the more it is used. As it gets bigger, it would push out the skull. So if you mapped the topography of the skull, you’d know from the outside what was going on inside. Gall called this process cranioscopy. Everyone else called it phrenology.
Although the scientific community didn’t ascribe to it, for over 100 years, phrenology was used by the general public to describe people. Using phrenology made you feel smart, educated and scientific. It also fit so well with people’s prejudices that it could and was used to discriminate against people they didn’t like.
Phrenology supported stereotypes that were cultural, national and racial. Something was wrong if your eyes were large or small, close together or far apart. You were lazy if your nose was too wide. You were conceited if your nose was too narrow. It was a one-stop shop for prejudice.
Correlations are still used today. You can’t randomly assign children to parents or people to geographical locations. But you can explore child abuse, obesity, diet, job advancement, leadership and the impact of television.