> I disagree. I believe it gives us a bayesian prior for the distribution of the population's attitudes. Sorry, I know I'm being obnoxious. The reason I'm real

> I disagree. I believe it gives us a bayesian prior for the distribution of the population's attitudes. Sorry, I know I'm being obnoxious. The reason I'm really sticking on the survey point is that people are very hard to survey. They have a lot of contradictory opinions ("I support banning preexisting conditions, banning lifetime caps on care, and banning junk insurance plans, even if it means I have to get insurance, like it says in the ACA, but I super hate Obamacare"), and are deeply, deeply ignorant. Here's a poll from 2010 where 41% of respondents couldn't name Joe Biden as VP in an open ended question [0], coincidentally also the poll that destroyed my faith in an informed citizenry. Tons and tons of thought goes into the way questions are phrased, because it has huge effects on outcomes, and if you want any kind of longitudinality you can never, ever change it (which is why we still have right direction vs wrong track despite its asymmetry). So if that's where we're starting, to pile on with a junk pollster asking an irrelevant question is like, I don't know, turning a hose on a drowning dog. Public opinion is very, very hard to discern, you certainly can't make blasé claims, no matter how much you really want to believe they obey Bayes' theorems (they definitely don't, see every election poll ever taken). [0]: https://www.pewresearch.org/religion/2010/09/28/u-s-religious-knowledge-survey-who-knows-what-about-religion/#Nonreligious https://www.pewresearch.org/religion/2010/09/28/u-s-religiou...