http://allendowney.blogspot.com/2014/02/correlation-is-evidence-of-causation.html http://allendowney.blogspot.com/2014/02/correlation-is-evide... Your comment

http://allendowney.blogspot.com/2014/02/correlation-is-evidence-of-causation.html http://allendowney.blogspot.com/2014/02/correlation-is-evide... Your comment is severely misguided in several ways. 1. Correlation does not prove causation, but it is evidence of a causal relationship. 2. Optimal decision theory doesn't require you to be 100% sure that evidence is causal. You can still use correlations to make good decisions. "Oh, it's not clear that eating paint causes me to be sick; it could be that my body getting sick causes me to eat paint." Doesn't matter! You still shouldn't eat paint. 3. History proved the fleeing middle class correct. Inner-city urban districts did go to shit. They were right, you would have been wrong. 4. Consider the four possibilities in the article I linked. One of them is vastly more likely than the others. * Poor inner-city immigration => increased crime (reasonable, with strong evidence) * Increased crime => poor inner-city immigration (violates temporal causality) * Secondary factor => (increased crime) ∧ (poor inner-city immigration) (reasonable, but we haven't seen strong evidence for this) * No causal relationship (unlikely in the face of all the data) You can't ask people to risk life and limb on the extremely remote chance that the world happens to work the way you wish it did, despite all evidence to the contrary.