You’ll quite often find something interesting (read: supporting your point of view) if you analyse a subgroup of an otherwise negative larger group. This one should be taught alongside the “screening for very low probability events” problem.