Links – 09/04/2017

  • A recent survey discussed by Eric Kaufmann here indicates that, in the West, whether one believes that it’s racist to want to restrict immigration in order to preserve one’s culture predicts whether one thinks immigration should be reduced. But I don’t even think it’s the most interesting aspect of the discussion, which includes lots of facts that explain why a platform that is explicitly pro-immigration is political suicide. In particular, the article mentions several studies which show that, on the issue of immigration, social desirability bias leads polls to vastly underestimate the opposition to immigration.
  • I have always been somewhat confused about the distinction between convergence in probability and almost sure convergence, because I never went through a book that offered a rigorous, measure-theoretic treatment of probability theory. I recently came across this excellent blog post by Matt Simpson, which explains the concept of convergence in distribution, convergence in probability and almost sure convergence. It’s the best explanation I have ever read on that issue and doesn’t require that you know any measure theory. On the same topic, I haven’t read this yet, but it also seems useful.
  • Quanta Magazine published a good piece on recent advances toward a physical theory of the emergence of life. Incidentally, it shows how computers are really changing science, by allowing scientists to do simulations and study the behavior of complex systems in a way that would have been impossible before.
  • Annette Gordon-Reed wrote a good article in the New York Times about how we should characterize Jefferson’s relationship with Sally Hemmings.  The bottom line is that, just because she was a slave, it doesn’t mean that Hemmings had no agency and we’re not doing her any favor by denying that she had. Gordon-Reed also wrote a book about Hemmings and her family, but I haven’t read it.
  • Chris Stucchio and Lisa Mahapatra published a great article in Jacobite where they criticize the way in which journalists use the term “bias” when they accuse artificial intelligence of bias against blacks, women, etc. They talk about COMPAS, a system that is used to predict whether a criminal will reoffend to assist legal decisions, which has been accused of bias. Robert VerBruggen also wrote a nice article on this last year, where he explains why this accusation is unjustified.