Religious service attendance and major depression: a case of reverse causality?