I think we need to discuss how unreliable the polls are and how the result of the election should probably end the pollster industry. They got it wrong on both Brexit and 2016 election. How did they get this so wrong?
Polls overestimated Democratic turnout. Trump got less votes than either McCain or Romney. edited: Obama - McCain 2008 69,498,516 - 59,948,323 Obama - Romney 2012 65,915,795 - 60,933,504 Clinton - Trump 2016 59,299,381 - 59,135,740 (maybe 100% counted)
Apparently there were a lot of Trump voters who were out of reach of the pollsters or simply were not inclined to reveal who they'd vote for. Probably a lot of the people who had reported they are undecided or in favor of a 3rd party candidate, in the end, shifted to Trump. Trump is like the "wildcard" candidate, and as such perhaps he had more appeal to undecided voters compared to a "you know what you're going to get" candidate like Clinton.
I was very clear their was an over sample going on. Many polls were saying the Dem turnout was going to be similar to the 2012 election. Nope!
Screw experience and judgement. I heard last night that the Trump supporters were not confident that Trump could deliver change but at least he would try.
I heard what I heard on GMA from a Republican spokesperson. Now there could have been record turnout in the Swing States that swung for Trump?
Yes, I also think once more of the wikileaks dropped people voted for a 3rd party (Johnson etc.) as well. Especially Once the realization set in that Clinton was basically selected and given special privileges during democratic debates. I'm also sure some people were just done w/ it all and didn't vote.
It comes down to this. More people were focusing on Nate Silvers accumulated model instead of individual polls. 538 made the mistake in letting the garbage results into his formulas. The independent vote indicator was largely ignored. Johnson crashed in the last week. I think most people realized Johnson was worse than Clinton and Trump. When polling undecided voters of this magnitude, it needs to be on a '1-10' scale of how likely someone will vote for each candidate vs an either or.
[Educational Post] It's now abundantly clear that the polls were nothing more than propaganda to suppress Republican voters by depicting a race that was out of reach. Yet another liberal dirty trick exposed in the most pivotal election in American political history. ABC News/Washington Post had Hillary up 12 points nationally as recently as a few weeks ago. NO CRED
This makes no sense. Most other poll aggregators had Hillary at 90+% chance of winning, while 538 had her 20 points lower. The real problem was all of the polls and not the poll aggregators.
Polls will change drastically from this point, but RocketsLegend you were right. I apologize for the dozens of times I called you wrong over the year. I'll man up and admit defeat now for real, you were the minority who were right.
Pretty much. Except this time the people were so pissed off they gave the media the finger and came out and voted anyways. The media and pollsters are no longer credible in any way. The folks have FINALLY figured that out.
I thin this is mostly right, but will point out aggregators like 538 grade and weight the polls, and they've perhaps given some polls a better grade than they deserved. I can't fault them too much though -- polls are graded on their accuracy based on history. The history is reliable with the status quo, but Trump's campaign shifted the map on who Republicans were courting making the election less predictable. I don't think this is any kind of existential problem for pollsters. The polling math will be updated with the new information of this and future elections and they'll get better at measuring the new normal.
