No-one doubts there is a correlation, but the notion of causality in “have an obese friend and your chances of becoming obese go up 57%” is junk science, according to real statisticians. Search Google Scholar for “The spread of evidence-poor medicine via flawed social-network analysis”.