General Discussion
Related: Editorials & Other Articles, Issue Forums, Alliance Forums, Region ForumsCNN poll is a "bit of an outlier"
according to Simon Rosenberg, a pollster who is frequently on O'Donnell's show. He is easy to find on twitter and always brings a good perspective. Not overly optimistic or dismissive of all polls that show cheato ahead, but good at presenting the big picture. As he says, there have been dozens of independent polls lately, with many showing Biden ahead and many showing recent gains for Biden.
Fiendish Thingy
(15,711 posts)Trump ahead of Biden with young voters by 11 points? Thats laughable.
The CNN poll did not disclose the size of the subgroups, including age, which is a troubling lack of transparency. What they did disclose is, for many questions, the MOE for 18-34 year olds ranged from 8 to nearly 10%. On many other questions, the sample size for 18-34 year olds was below 125 respondents (I guess some respondents didnt answer every question?), and so no percentage was reported for that question.
senseandsensibility
(17,224 posts)Rosenberg also points out that the CNN poll is contradicted by the recent CBS poll.
WarGamer
(12,509 posts)Meadowoak
(5,574 posts)LetMyPeopleVote
(145,911 posts)senseandsensibility
(17,224 posts)how CNN should include the fact that this poll is an outlier in their coverage. It's one of the things that makes me think they have an agenda.
WarGamer
(12,509 posts)Reality is... the election will be close.
Better turnout will win and the side that keeps the best news stories in the news at the end... will win.
Anyone predicting a blowout is lying or doesn't follow politics.
W_HAMILTON
(7,878 posts)...because it does not suggest the election will be close: it shows Trump winning by 6%.
Sympthsical
(9,193 posts)And get a little bored with, "Polls are always wrong!" whenever someone sees a result they don't like.
However, looking at the data of this one . . .
Yeah, this poll is trash, lol. Straight up throw it in the bin. The sample size, the margins of error. I just can't see how any useful data or insight can be gleaned from the methodology of this one.
At present, the election is closer than is comfortable. We'll see what materializes in the beginning of October to get a real sense of how things may go.
But I'm not sure I would even report this poll as a journalist. It's so incredibly wonky, it provides no value outside of headlines.
Oh . . . I've just answered my own question there.
dsc
(52,173 posts)that said, a poll being an outlier may mean it is a leading indicator of a change in the race or that the poll is wrong. Time will tell which is the case here.
Tribetime
(4,730 posts)Is to make up for the electoral college
ITAL
(654 posts)He had a popular vote win of about 5.5%. Some pollsters I've read seem to think the margin doesn't need to be 5+ to guarantee a win this time around because of changing dynamics of the various swing states (more like 2-3), but I guess no one really knows.
Polybius
(15,528 posts)Celerity
(43,762 posts)have won (and via a nightmare scenario, a 269-269 EC tie, with an illegally gerrymandered (in a handful of close states) rigged US House Rethug state delegation advantage electing him). Due to gerrymandering in WI, FL, and NC, it is impossible to get them under the 26 delegations needed, and we may perhaps lose some more state US House delegations that we are tied or have narrow control of, in terms of districts held by Dems atm.
But as i said, I've read a few analyses that suggest the swing states are now trending closer to the popular vote than they were four years ago.
Celerity
(43,762 posts)I do not see an on balance Biden advantage of 4.5 per cent or greater in the key swing states (GA, AZ, NV, MI, WI, PA, NC) when the entirety of the data is analysed.
Maybe in NH. Maybe in NE-2 but there are no recent polls there. Maybe in VA. If we lose those, then the trend nationally that would portend is dire for us.
Deffo not in FL (which is now a pretty red state, despite some saying (incorrectly IMHO) it is absolutely in play). Deffo not in ME-2. Deffo not in TX , which is the true tell (claiming we will win it) of a non serious prognostication for decades now.
The best thing for us atm crazily appears to be RFK Jr's negative effect on Trump more than Biden.
Here is my list of the things that most likely can hurt Bidens chances the most (in no order other than the first):
The price inflation for key things (food, petrol/energy, rent, etc) gains steam as we near the election, with aid from giant corporations and bad international political actors who desire a Biden loss.
Israel goes even more wild and the Biden administration protests but does not stop military aid.
The TikTok ban (instead of a ByteDance sale) becomes more certain.
The southern US border situation becomes far worse.
Biden opens up more areas for drilling and/or increases leasing, especially if they once again falsely claim that they are legally forced to. They falsely claimed that very thing with their Gulf of Mexico oil lease auction, but then their own DoJ filings showed that they were not. Many of my own personal social set (mostly 18 to 36yos, I am 27) who are the most engaged on climate change still talk about that.
Trump is not convicted in any trial before the election. Even worse would be a mistrial in NY that happens in a way wherein double jeopardy law prevents a retrial.
RFK Jr wnds up drawing more equally from both Biden and Trump, and also West and Stein do not fade much and pull 4 to 6 percent or so combined in key swing states. Another wild card would be RFK Jr drops out and endorses Trump.
There are other dangers, but that is enough for now.
Also, Trump has huge potentially disatrous issues as well that could (and hopefully do!) ratfuck him hardcore.
Time will tell.
Polybius
(15,528 posts)7 isn't needed, but it would be nice.
MFM008
(19,836 posts)It was conducted online
Second.... land lines.
Every time Biden starts to inch upward. CNN rushes in with a poll to bring it all down again.
senseandsensibility
(17,224 posts)And as Simon Rosenberg says, they don't put their outlier results in context.