Things like social sciences employ statistics from non-reproducible data involving an immense number of variables that often go unaccounted for. These are then used to make political (or other) claims that are dubious. Someone like Michael Moore is guilty of this all the time, when he’ll use statistics to say something like, states with the death penalty have equal or more homicides than states without it, so the death penalty does nothing to deter homicides. There’s something fundamentally flawed about the argument because it doesn’t take into consideration variables like unemployment, income, religious faith etc. Sam is also guilty of this when he says countries that are statistically more atheist—Sweden, Netherlands, etc.—have statistically lower homicide rates, therefore, religious morality is not a necessary deterrence to people committing homicides. I’m not arguing for or against this, I just think it’s flawed and undermines good science/reasoning/argumentation that can be replicated more consistently and empirically with crucial variables accounted for. This debatable science is applied to everything from global warming to weight loss. I’m not a fan.