A large organization can usually identify groups of "similar enough" incidents to identify the correlation. (Can the numbers prove causation? Not to the standar

A large organization can usually identify groups of "similar enough" incidents to identify the correlation. (Can the numbers prove causation? Not to the standards of a scientific journal, but you can't exactly induce 100 incidents to do a controlled study.)