2016 Postmortem
Related: About this forumPOLL: North Carolina HighPoint University: Clinton +1 with leaners, +2 w/o
Likely voters with leaners:
Clinton 43, Trump 42, Johnson 10
Likely voters w/o leaners:
Clinton 42, Trump 40, Johnson 8
http://www.highpoint.edu/src/files/2016/09/47memoA.pdf
404 likely voters, landline and cell phone, conducted 9/17 - 9/22
4.9% MOE, sample weighted based on 2008 and 2012 demographics
DemocratSinceBirth
(99,708 posts)That suggests some of the swing states will be on a knife's edge...
Loki Liesmith
(4,602 posts)DemocratSinceBirth
(99,708 posts)Can you think of a more self defeating emotion?
Even if I was so scared I was fouling myself I wouldn't let my enemies know.
My buddy was in the freaking Battle of Fallujah. I asked him if he was scared. He said no because it wouldn't do any good and only increase the likelihood of getting injured or killed.
Loki Liesmith
(4,602 posts)R friends are always confident, never freak out at polls.
What drives this?
DemocratSinceBirth
(99,708 posts)He had boundless empathy , boundless physical courage, and boundless moral courage.
BTW, is your model based on nat'l polls, state polls or a hybrid?
Loki Liesmith
(4,602 posts)I use 2012 turnout numbers to estimate a national number and validate the runs based on the similarity to the current national polling.
My last number had a spread of 2.45 advantage Clinton. That looks fairly likely.
Can't run it now as work firewall prevents me from downloading polls from HuffPost. I expect today will put a dent in Clinton's numbers.
DemocratSinceBirth
(99,708 posts)Nate is getting a lot of push back on his model . It is is not as if those questioning it are nobododies; Sam Wang, David Rothschild, and NassimTaleb are not nobodies.
Loki Liesmith
(4,602 posts)I said that Nate Silver's model is almost entirely correlated with the national numbers. I regressed his national numbers against the moving average from huffington post and found that the national poll average explains around 60-70% of the variation in Nate's estimates of probability.
I said I could probably do almost as well as Nate just converting the national poll average into a direct probability.
but my model uses medians like Sam Wang and extrapolates missing data from trends.
DemocratSinceBirth
(99,708 posts)The same eye test that informed me of the winner, at this moment, in every election since 1976.
To be candid I don't precisely remember who I thought would win in 00.
Loki Liesmith
(4,602 posts)I was actually a Nader advocate, until I saw what might happen. I switched over to Gore. Hope I didn't do to much damage.
DemocratSinceBirth
(99,708 posts)molova
(543 posts)Cheer up and rec.
Loki Liesmith
(4,602 posts)All battlegrounds tightening.
could lead to a breakout either way.
CajunBlazer
(5,648 posts)In addition, on Nate Silver's website he has corrected this poll for known bias - went from Clinton +1 to a tie, and he has assigned it a relatively low weighting for accuracy.
smorkingapple
(827 posts)And it's why Plouffe and Messina are very confident of a Clinton win.
CajunBlazer
(5,648 posts)geek tragedy
(68,868 posts)1) that's not necessarily a warranted assumption;
2) exit polls aren't necessarily correct in terms of demographics--they can easily undercount white voters, or nonwhite voters
Dawson Leery
(19,348 posts)triron
(21,984 posts)new early/absentee voting info?
triron
(21,984 posts)triron
(21,984 posts)have updates on ballot requests by party?