2016 Postmortem
Related: About this forumAllison Grimes ahead of McTurtle 42% vs.40% from her internal polling
Same polling that called the race for Harry Reid when every other poll called Reid a goner.McConnell according to this poll has a 58% disapproval rating in Kentucky.If those numbers are correct he's DONE!! no way an incumbent can get reelected with that kind of disapproval.The Mellman polling firm sticks by there numbers saying we are confident we are accurate http://www.dailykos.com/story/2014/10/01/1333608/-KY-Sen-Mellman-Group-Poll-Has-Grimes-D-Leading-McConnell-R-42-40
WI_DEM
(33,497 posts)it certainly was a big polling sample. We need to try and defeat somebody along with Roberts and I think the best bets are KY & GA if Dems can GOTV.
bigdarryl
(13,190 posts)They don't use the Likely voter models as every other polling outfit does
Arkana
(24,347 posts)AnalystInParadise
(1,832 posts)People want a pony really badly, but internals that show any candidate barely ahead or tied is not exactly good news. In-house bias is usually worth 4-8 points.
liberal N proud
(60,334 posts)If this holds, this will be huge!
Drunken Irishman
(34,857 posts)Never trust internal polling. They're there solely to garner positive attention for a campaign and generally done by the losing campaign to indicate to supporters they still have a chance.
I see this all the time locally because our Democrats in Utah are almost always non-competitive in races and yet, they'll release internals showing the candidate down by only five-to-eight points.
Then they'll lose by 20.
bigdarryl
(13,190 posts)Drunken Irishman
(34,857 posts)You can dismiss it, Darryl, but it's reality. I've worked on losing campaigns. I know how the internals work.
They're generally done using demographics that are favorable to the candidate and only leaked when they need to recreate buzz and keep donors and supporters still believing so that the money keeps rolling in and people still get out the vote.
Most winning campaigns will not release internal polls as they don't need to. If a campaign is releasing an internal, I guarantee you they're losing.
FBaggins
(26,740 posts)Internal polling is done far more frequently than media polls... so you tend to see plenty of variability in results even when the race isn't changing.
But only the ones at the far end of the distribution ever get released to the press.
SmittynMo
(3,544 posts)(spoken is a soft, slow voice)
Ummmm Yuuup, I'm losing. What is a turtle to do?
BTW Jon Stewart does him great!!!
Proud Public Servant
(2,097 posts)Mike Daniels
(5,842 posts)GGJohn
(9,951 posts)is really doing, it's more of a pep talk to the persons campaign staff and donors.
Sorry to be a wet blanket, but that's just the way it is.
Dems2002
(509 posts)As someone who worked extremely closely with a polling firm for ten years on over 100 polls, I'm sorry to say that using registered voters versus likely voters is an inherently flawed approach, particularly in a non-Presidential election year. Turnout is likely to be under 50% and stratified random sampling that takes into characteristics the age, demo, party affiliation, gender, location of these voters is the only way to come close to accurately predicting the result.
My guess is Allison is losing by at least 5% if not more. Not to say that I'm not rooting for her and wouldn't love for this poll to be accurate, but it's just not the right way to conduct polling to predict election results.
WI_DEM
(33,497 posts)while registered voters showed Obama winning.
That was an internal poll, and what they did is basically the opposite. They created a special likely voter universe not based in reality, but rather the type of universe Romney would require turning out to win the election. So they under sample blacks, young voters and Democrats and oversample middle age white men.
Almost all Republican polls conducted in Presidential Election years do this in order to have a chance of putting their candidate in the lead.
Every polling firm has their own likely voter universe model. The ones who are unbiased can do a pretty decent job of coming really close to predicting the results w/o their margin of error. But the goal of internal polls is to show your candidate winning, or capable of winning the race in order to attract money. W/O money, there's not a chance in hell.
Mutiny In Heaven
(550 posts)The people of Kentucky simply prefer Mitch McConnell and his brand of politics. He's the type of person who is most palatable to people there, much like Abbott over Davis in Texas, another race which is a write-off despite some fringe suggestions to the contrary.
rock
(13,218 posts)Still I like being ahead by any amount and in any poll.
Addendum: except, of course, the "Are you hated more than Hitler?" poll.
a kennedy
(29,663 posts)AnalystInParadise
(1,832 posts)Mellman is the ONLY poster showing this race less than 4-5% for Mcconell.
Mike Daniels
(5,842 posts)Eric Cantor.
His internals showed him running away with his election and we all know happened in the end.
Meanwhile this internal has a 2 point lead. I wouldn't be banking on Grimes being even slightly in the lead based on one internal poll.
TheKentuckian
(25,026 posts)though and more volunteers are needed to help push.
I tend to believe it is close.
Corey_Baker08
(2,157 posts)SO LETS NOT RELAX, LETS NOT GET COMFORTABLE, LETS NOT ASSUME THESE POLLS ARE CORRECT. I WANT, & WE DESERVE THAT KENTUCKY DEMOCRATS GO TO THE POLLS AS IF THIS ELECTION IS THE DIFFERENCE BETWEEN LIFE OR DEATH....