Any time stats are used as "evidence" then you know its bs. You can just keep picking variables until you got a chart you want.