If you're on HN, chances are, you're not very correlated to the sample from this research study... which looks entirely at correlation, not causation. I'm interested in a long-term study that identifies real differences in health outcomes, not just population bias.