hsrob on 31 March 2009
Just a quick note regarding the whole correlation versus causality thing.
In complex systems that can't be reproduced easily or replicated in a lab (and hence difficult to prove causation), the Bradford Hill criteria can be used.
They were initially used in medicine apparently but are more generally applicable. There are about 8-10 criteria which help to describe the minimum conditions required for a causal relation between two things.
It would be interesting to see if it could or has already been applied to the above situation. (Brady Act and the Nonfatal firearm incidents data)
I'm going to have lunch.
