Correlation and causation are terms that describe relationships. Unfortunately, they are easily confused. In causation, one action or occurrence causes another: He ate and so his hunger went away. In correlation two things occur together, but one does not necessarily cause the other: He ate, and it was noon.
If one occurrence causes another, there is generally correlation as well as causation: Smoking causes lung cancer, and lung cancer is associated with cigarette smoking. However, correlation without causation is quite common.
In simplest terms, correlation means things occur together. Summer has warm temperatures and winter is associated with cold. Children get taller with age, as a rule.
Correlation is not hard to find. Statistics demonstrate, for example, a correlation between maternal medical care and infant health. However, they do not prove that maternal health care causes healthy babies. Perhaps mothers who visit their doctor also watch what they eat, exercise properly, and avoid alcohol and tobacco. Perhaps they are in better shape financially, and thus have less stress. Perhaps they live in a country with good public health practices and are thus exposed to less disease. It takes more than a match to prove anything besides correlation.
The crucial point about correlation, therefore, is that it cannot prove causation, though it often indicates to researchers where to look for it.
Causation means one thing causes another, a simple enough idea. However, sometimes there can appear to be causation when a relationship is really correlation. To use a famous example, children at birthday parties sometimes get wild. The supervising parents then agree that the children’s behavior proves that sugar makes them hyperactive.
However, the parents are wrong. Parties themselves may make children unruly. The atmosphere and the expectations are very different from everyday, and this holiday feeling may itself cause boisterous behavior. Who knows?
It is true that repeated controlled studies have clearly indicated the lack of a relationship between sugar ingestion and activity levels in children.
Controlled studies are one way that causation can be proved. In such a study’s simplest form, two similar groups are evaluated. They are exposed to the same experiences, except for one variable.
For example, two groups with incipient colds might receive lemonade each day, but one group would have a particular medicine added to the drink. If their colds measurably improved or worsened, that would actually show causation.
Something that someone “just knows” is not causation.
However understanding correlation and causation often runs counter to “common sense.” For example, vaccination and the diagnosis of autism have both grown more common in recent years. In addition, autism is often diagnosed at an age when children have recently had certain vaccinations. Therefore, many parents believe that vaccinations cause autism.
They are wrong. Many experiments and much epidemiological data have actually shown that vaccination prevents dangerous diseases, while it inexpensively protects entire communities from epidemics. However, many parents still go to great lengths to avoid protecting their children, because they confuse correlation with causation.
Correlation means that two occurrences are associated, like going to the movies and eating popcorn. Causation means that one causes the other, like smoking cigarettes and getting wrinkles.