A retrospective study, the kind of study that notoriously cannot distinguish causes from effects. The fact that the subjects are drawn from an existing populati

A retrospective study, the kind of study that notoriously cannot distinguish causes from effects. The fact that the subjects are drawn from an existing population who live in high-pollution areas is not, and cannot be, eliminated as a cofactor. In short, do these people have diabetes because they're exposed to pollutants, or is the state of their health and their place of residence both connected to some third unexamined factor? The study cannot tell us.