In the sense used by political statisticians, ecological association is the association between two characteristics both measured at the aggregate level rather than the individual level, and has nothing to do with ecology. For instance, it may be shown that regions with high mining employment also display a high vote for the left-wing party. The ecological fallacy is to infer from this that miners vote for the left-wing party. From the given facts, nothing is known about the individuals in the region and there may (although in this case there is unlikely to) be some quite different reason for the association than the obvious one.




