Earlier work on this topic tended to reflect the biases of research disciplines: psychiatrists looked at antidepressant drug use; economists looked at the unemployment rate; and sociologists and demographers emphasized factors such as family structure, religious composition, and age structure.