It makes Hillary sound like Lord Vetinari. Which would be awesome, and just makes me want to vote for her more. Of course, Vetinari would never actually let himself be subject to an election that he wasn’t already certain of winning in advance.
So The Donald gives his “My First 100 days” speech prepared by his handlers in a desperate last-ditch Hail Mary gambit, which is widely ridiculed even as another woman comes forward with accusations of sexual molestation, and Hillary jumps to a 12-point lead. Maybe The Donald’s next trick will be to throw a tantrum and threaten to hold his breath until he turns blue if the polls don’t improve … IIRC, the ABC poll has historically been one of the most accurate predictors …
Hillary Clinton has a 12-point lead over Donald Trump and has reached 50% support nationally among likely voters, a new ABC News tracking poll shows.
The poll shows Clinton with 50% support to Trump’s 38%, with 5% backing Libertarian Gary Johnson and 2% supporting the Green Party’s Jill Stein.
I voted Friday. Straight D ticket. Except for when there were two R’s locally running unopposed. I wrote in Gilbert Gottfried.
Too funny, silenus. 
The New York Observer, by the way, is owned by Trump’s son-in-law. And it is one of those papers which you can get free with your spare airline miles.
Rex Reed still reviewed movies there, for Og’s sake.
Trump is now just suggesting his fans go commit voter fraud for him:
[Quote=Loser Donald]
Maybe they’ll vote for Trump, I don’t know, maybe I shouldn’t be saying that. I may be hurting myself, you’re right. You’re right. Maybe they’re going to vote for Trump. All right, let’s forget that. It’s okay for them to do it.
[/quote]
Sad!
Glad to see Hillary even further up in the polls. I canvassed today on Cleveland’s East Side, urging people to vote early. Only spoke to two actual voters in two hours, but dropped off Hillary lit at a lot of nobody-home-right-now addresses.
Those voting absentee, make sure you put enough postage on your envelope. It’s becoming an issue here in Ohio. A single First Class stamp isn’t enough.
True, but the American political system has changed a lot since 1968. We see so much more of our leading candidates over very long campaigns now, and in a 24/7 news cycle. I doubt any politician will ever again be able to capture the Democratic or Republican presidential nomination, lose in November, and then win the nomination again. These days, I think you get one chance.
538’s Now-Cast has Iowa for Trump, even though both the Polls-plus and Polls-only have it for Hillary.
Seriously, the more I look at the WikiLeaks emails, the more admiration I have for the way her campaign is run - thoughtfully and carefully, with approaches hashed out and debated, and occasional blunt discussion. Thanks, WikiLeaks!
We can add a +1 to “things big storms are good for”.
It feels like he realized that running on his personality wasn’t going to win, that in a mudslinging contest it’s really hard to beat “sexual assaulter”, and felt like he had to pivot towards policy to save his ass… Not realizing that Clinton has been beating that drum since day one, and his policies are shallow bullshit that anyone with half a brain can see through.
I am, alas, exposed to a few right-wing Trumpeters on social media, and the thing they’re hollering about right now is how the recent ABC poll (showing Clinton up by 12 points) is “oversampled.” They point to this paragraph in the poll’s fact sheet:
METHODOLOGY – This ABC News poll was conducted by landline and cellular telephone
Oct. 20-22, 2016, in English and Spanish, among a random national sample of 874 likely voters.
Results have a margin of sampling error of 3.5 points, including the design effect. Partisan
divisions are 36-27-31 percent, Democrats-Republicans-independents. (emphasis added.)
I can’t tell if the poll result takes that odd sampling – 33% more Dems than Reps – into account. Is that what they’re calling the “design effect?” I’m no expert in the nitty-gritty of poll sampling, but it does seem odd that they contacted so many more Democratic voters. Can anyone here elucidate? Thanks!
In re Wikileaks:
Yes but he hits the key GOP essentials, gut the federal government as much as possible and block liberal supreme court nominations. It will probably be enough to encourage some Republicans that were disgusted by his antics to hold their noses and vote for him.
Although the anti-free trade stuff is an odd mix which will turn off some of the corporate republican types.
“Design effects” should actually be called design defects, at least if the intent of the poll was honest. They’d be things like where exactly are you calling (areas with mostly college students? Areas with heavily Hispanic populations? whiter-than-thou suburbia?) or at what times (what kind of people are likely to be willing to take it).
Also, what do you do when you hit an invalid point. A US political poll which happened to get someone not eligible to vote shouldn’t count them as “undecided”, for example.
The idea is to design the polls to call, not just 1000 people, but 1000 people who actually happen to be representative of the population being investigated.
Some twenty years ago there was a poll which said “98% of Basque want independence” and was published in all newspapers with a large amount of exclamation marks. They had polled 2000 people, all of whom lived in a township where the last electoral results were those. The response from others was “now ask 2000 people in, say, Monteagudo” (that is, in a location where if you say you’re an independentist people will actually move over to stare at you and verify that you do look like a normal person). If the poll designers wanted to find out that Basque people want independence well then their poll was beautifully designed, but if they wanted to know the actual opinion of the actual Basque population they should have cast a wider net.
I’m no statistics genius but will weigh in on the theory that if someone comes along and corrects me, I will learn something and probably enjoy the lesson, since statistics is, at least in limited doses, kind of fun.
First - it seems misleading that you (or the Trumpies perhaps) are characterising this as “33% more” Dems. I know that 36 is a third larger than 27, but when written this way it sounds as if the make-up of the people being polled was something like 10 percent Republican and 43 percent Democrat. It’s not THAT unevenly distributed. Rather, 9 percent more of the total sample was Democrat rather than Republican. Perhaps this simply reflects the pool of voters they polled, rather than suggesting that they went in deliberate search of extra Democrats.
As to the “design effect,” it’s been more than 30 years since I took statistics, so I don’t recall the meaning of that off the top of my head. When I took a glance at the Wikipedia entry, though, I’m pretty sure it is beyond the ken of most everyday voters (I’ll be charitable and not single out Trump supporters as being less likely to fathom it, especially since I personally was not willing to slog through the math).
Having said that, I assume “the design effect” has to do with the fact that the make-up of a sample may not be identical to the make-up of the overall voting population whose behavior you are trying to predict.
For example, if the entire voting population is (using totally made-up numbers) 35 percent Democrat, 40 percent Republican, and 25 percent Independent, but your sample turns out to be 38 percent Democrat, 39 percent Republican, and 23 percent Independent, and it is very clear that party affiliations correlate strongly with how people vote, then you will want to adjust the results of your poll so that you compensate for the difference between the make-up of the group polled and the make-up of the electorate.
I welcome correction from more knowledgeable posters - maybe all I did above was to state the obvious, without actually touching on what the design effect does?
It’s funny that you say this, because when I first typed up my post, I wrote “9% more.” Then I realized that wasn’t technically accurate – 36 isn’t 9% more than 27. I thought that using “9%” would be misleading because it would understate the gap, so I changed it to 33% to be LESS misleading. Ah well. Just goes to show you can’t please everyone. 
Two weeks to go. Anyone think the media’ll try to make this a horse race one last time?
“TRUMP SURGES TO WITHIN 2 POINTS OF CLINTON”
“TRUMP: LAST-MINUTE LEAD IN SWING STATES”
etc. etc.
I know, right? I just finished editing a paper for a client that had some calculations in it - while the mathematical specifics were a little different, the situation was roughly analogous; there was more than one way to characterise the data. I preferred it the way economists do the calculations, she preferred it the way financial analysts do. Since it was her paper, she got to do things her way in the end - but I think we were both disappointed that we could not convince the other person of the rightness of our approach.
Nah, I think the narrative shifted to “TRUMP CAMPAIGN DYING” and “TRUMP THE PERVERT” after the Billy Bush tape came out, and I doubt it’ll shift much between now and election eve.