> We absolutely have statistical and theoretical frameworks for evaluating collective behavior. Yes, we do, but their predictive power, if it exists at all, is

> We absolutely have statistical and theoretical frameworks for evaluating collective behavior. Yes, we do, but their predictive power, if it exists at all, is very poor. Science is about building models with good predictive power. If you don't have that, you don't have a science, at least not in any way that matters.