All you really need to survey the U.S. population is about 400 people for a +/- 5% error rating at 95% confidence.
Properly randomized in one area or across the country cause thats part of my issue. Results in this state overall would be drastically different than say california.
Once again still dont care for it as its millions of other people that could have answered very differently. Polls have ended up being wrong in the past so I have no reason to trust them.
It was randomized throughout the country. Believe it or not people doing statistics know more about it than someone who has clearly never taken even an intro stats course does.
Also no real polls showed Trump having no chance. People saw a 28.6% chance of winning (in a forecast derived from polls, not a poll) and assumed that would mean he would lose for no reason. https://projects.fivethirtyeight.com/2016-election-forecast/
28.6% chance means in over 1/4 of simulations they ran Trump won. It turns out reality conformed to that 1/4.
