Refraction (Bent Thinking) #53

Correlation Proves Causation (cum hoc ergo propter hoc) : a faulty assumption that correlation between two variables implies that one causes the other. Refutation: statistical correlation is not causation. The only way to prove causation is to repeat the event and remove all other possible causes. If the effect still follows, there is causation. Example: Ann: Bob, in this last recession, there have been six new churches start in our city. During the same time period, there have been 5 new taverns opened up. Therefore, religion drives people to drink!  Side Note: Statistical correlation as proof or even suggesting proof is one of the more insidious and common fallacies in science and public policy. Always demand proof of causation!