Andrew Gelman writes: Social scientists who use medical analogies to explain causal inference are, I think, implicitly trying to borrow some of the scientific and cultural authority of that field f…