Originally Posted By ecdc Why trust Nate Silver, he's only an expert.... Silver went through this in 2010 with some Democrats when he predicted a fairly big loss in D seats in the House. He was, of course, correct. Could he be wrong? Could he be off? Of course, but I have yet to see a compelling argument that he is.
Originally Posted By andyll <<its not a MYTH of oversampling democrats. Its is precisely why Kerry was ahead in so many polls by 4 points heading into election day.>> Kerry was not ahead. You are falling into the trap of only paying attention of what polls are reported in the media. What Silver and electorialvote.com does is aggragate the polls. Silver takes it beyond that. Based on historical data he will weight polls that tend to oversample. ElectoralVote in 2004 had it dead right the Sat before the election: <a href="http://www.electoral-vote.com/evp2004/Pres/Maps/Oct30.html" target="_blank">http://www.electoral-vote.com/...t30.html</a> Silver goes into detail about his methodologies and why his results are the way they our.
Originally Posted By Dabob2 It happens quite frequently and definitely isnt a MYTH." It is definitely a myth, and Silver shows why. When you talked about "the polls showing Kerry leading" I could see where the problem is. You are getting your info from unreliable sources. This whole "They showed Kerry leading right up to election day!" thing is itself a right-wing myth. There may have been an outlier or two, but Silver shows that the aggregate of the polls at the time showed Bush leading by just about the amount he won by. But the far right loves to feel oppressed and conspired against. So they use an outlier poll and say "aha! See, we told you the polls are biased in favor of Democrats. " And if you get your info from unreliable sources who will cherry-pick in this way and don't view it critically, this is the result. You have to read or view various sources, including those that aren't so myopic about US politics (BBC America is good) or are at least mostly apolitical. They're getting harder to find but they exist. Silver himself, though he deals in politics, is himself a basically apolitical number cruncher. The right was fine with him in 2010 when he showed them winning big, but now that he shows Obama winning, they claim bias. This is itself bias, if you'd only see it.
Originally Posted By Goofyernmost I did a poll on the internet this year. They asked me who I would vote for and about many other topics that seemed legit. At the end, they showed the results of the poll and it showed Romney with 86% of the popular vote and Obama with 12%. My goodness those polls are accurate aren't they? Ask Truman how accurate polls are. The only results that count are the number of votes cast...nothing else.
Originally Posted By Dabob2 True... but the polls have proven to be remarkably accurate in recent years... 1948 was a long time ago. And Silver doesn't include fly by night polls in his conglomeration.
Originally Posted By skinnerbox Polling needs to change with the times. Social media is now far more important for communication than talking on the phone. And polling only land-based telephones will skew the numbers way beyond basic voting demographics. Nate is doing his data analysis correctly. Anyone who relies only on one or two polls, especially those conducted with small sample sizes, land-based phones only (robopolls), or mostly during standard working hours, will get essentially worthless results. And it's finally time to dump Rasmussen as a 'reliable' polling organization, for just these very reasons.
Originally Posted By planodisney Rasmussen and pew were the only 2 polls to get the 2008 election spot on, however Rasmussen was graded top because its polling was more consistant. this from multiple sources and most notably a Fordham study after the elction. Overall, Obama averaged almost 1.5 points higher in all polls taken and recieved a higher number in 17 of the 23 polls than he ended up with in final election results.
Originally Posted By planodisney One thing I realized is that Im getting pre-election polls and exit polls mixed up concerning the Kerry vs bush election. Most of the media were informally projecting a great night for kerry based on exit polls. If anything, after looking at how accurate polls have been over the last few elections, I would feel very comfortable if I were Obama and extremely nervous if Romney.
Originally Posted By skinnerbox Rasmussen does only robo-polling, which restricts it to land-based phones only. It's against Federal law to robocall a cell phone. One-third of American households do not have a land-based phone line. And that number has been steadily increasing with each passing year. Far more households had land-based phones in 2008 than they do now. Therefore, Rasmussen's polling of only land-based lines is now outdated and heavily biased, since most of land-based lines are used by households comprised middle-aged residents, many of them retirees. And many if not most of those retirees tend to vote Republican. It doesn't matter if Rasmussen was correct four years ago, because the demographics for land-based phone usage has dramatically altered in just the past few years.
Originally Posted By planodisney And yet their poll is basically right there where everybody else has it. They have it at 48 to 48. But your point is valid. my wife and i dont have a land line anymore. We strictly use cell and we use skype to keep in touch with her humongous family in Brasil.
Originally Posted By skinnerbox <<And yet their poll is basically right there where everybody else has it.>> Take a look at what Nate Silver is projecting. Nate has been the most consistent statistician for political races for years now, being correct over 90% of the time, even during 2010 mid-term elections. He's not calling it a tie, not by a long shot. You're simply projecting what you want the election to be, not what it's actually shaping up to be.
Originally Posted By andyll <<Rasmussen and pew were the only 2 polls to get the 2008 election spot on, however Rasmussen was graded top because its polling was more consistant.>> Rasmussen was wrong in 2008 on Indiana. (most were but Rasmussen was way wrong) In 2000 & 2010 they were horrible. In 2004 they got florida right but thats only because they were so biased on the Republican side. The other pollsters were more correct in FL because they were predicting a slight Democrat lean while Rasmussen predicted a blowout. Silver (amount others) has studied Rasmussen and they showed a huge bias towards republicans in the last 3 weeks of an election. <a href="http://fivethirtyeight.blogs.nytimes.com/2010/11/04/rasmussen-polls-were-biased-and-inaccurate-quinnipiac-surveyusa-performed-strongly/" target="_blank">http://fivethirtyeight.blogs.n...trongly/</a>
Originally Posted By Dabob2 "If anything, after looking at how accurate polls have been over the last few elections, I would feel very comfortable if I were Obama and extremely nervous if Romney." There you go.
Originally Posted By planodisney skinnerbox, Im not an idiot. Im not projecting anything to be the way I want it to be. What good would that do? Im simply calling it like it is when it comes to NATIONAL polls. It is deadlocked in most polls. The battleground polls are a different story. You can call me or anybody else any kind of name you want, but will you for petes sake stop pretending to know who I am and the way I think. You have no clue whatsoever about either. Can you ever respond to anyone without being condescending or a jerk?
Originally Posted By skinnerbox <<Im simply calling it like it is when it comes to NATIONAL polls. It is deadlocked in most polls.>> And what effn difference do national polls make for our Presidential elections? Answer: NONE WHATSOEVER Honestly, plano, why even focus on this? Who cares what the national polls are claiming about the race being tight. We have never elected a President in this country by popular vote. We've always been saddled with the electoral college, so national polls are worthless. Completely worthless. So why do you even bother paying attention to them, or bringing them up on the boards? <<The battleground polls are a different story.>> Exactly! And that's what you should be spending time on, not some pointless collection of statistics that doesn't matter a hill of beans, that only serve to gin up the horse race that the media has been trying to create. <<You can call me or anybody else any kind of name you want, but will you for petes sake stop pretending to know who I am and the way I think. You have no clue whatsoever about either.>> You have been acting as someone who wants others to believe that this is a very close race, simply by the nature of your posts. Since you've already made previous statements in support of Romney, which may or may not still be true, it's easy to assume that like most Romney supporters in the media, you would naturally gravitate towards those polls which would give you hope that he had a good chance of winning. The national polls do that. The swing state polls do not. You obviously chose to focus on the national polls, for some reason. I'm just extrapolating from what other Romney supporters have done by focusing on the very close national polls, and applied their reasons to you. If that offends you because it isn't true, and it's not the reason for why you're focusing on the national polls, then I'm sorry. <<Can you ever respond to anyone without being condescending or a jerk?>> According to other conservatives such as yourself, I seem incapable of doing that, at least with folks like them. So, let me correct your original statement to make it more applicable: "Can you ever respond to any conservative without being condescending or a jerk?" I've tried, but it's really, really difficult. And truth be told, I don't really want to make that much of an effort, since they're bound to hate me no matter what I say or how I act, simply because I'm a liberal.
Originally Posted By DDMAN26 <a href="http://www.politico.com/blogs/media/2012/11/the-redskins-rule-favors-romney-148367.html" target="_blank">http://www.politico.com/blogs/...367.html</a> Good news for Romney bad for Obama?
Originally Posted By SingleParkPassholder The criticizing of Silver has begun (got this from DDMAN's link): <a href="http://www.politico.com/blogs/media/2012/10/nate-silver-romney-clearly-could-still-win-147618.html" target="_blank">http://www.politico.com/blogs/...618.html</a>
Originally Posted By DyGDisney Post #56 --- perhaps this is what undecided voters have been waiting for?!
Originally Posted By Dabob2 <The criticizing of Silver has begun (got this from DDMAN's link):> Oh, it began a while ago (if you get my nutty uncle's emails). Most people who criticize Silver don't understand his percentages and what they mean. If he has Obama at 75% probability to win (as in that article; it's now up to 85%), that doesn't mean that he's going to win 75 or 85% of the vote. It means that based on what the polls in the various states have shown at various times in the cycle historically compared to actual election day numbers, one candidate or the other has x-percentage chance of winning that state. You add up the states, and get the overall probability of winning the electoral college and thus the election. Silver himself has the popular vote going to Obama by just 2 points right now. Yet an 85% chance of winning the electoral college. That seems like a contradiction, but it's not. Silver points out that with the leads Obama has in various states at this late date, it would be unusual indeed if he were to lose. There has been TONS of polling in the swing states, and Obama has been winning around 18 out of 20 polls in those states in recent days. Silver points out that they COULD all be wrong, or all biased in the same direction (despite their different methodologies) but that's unlikely. His model figures there's about a 15% chance of that. So it could happen. It just hasn't ever happened for 32 years, and polls are much better now (and more numerous, which should increase their reliability when taken as a whole - larger sample size and all that). Interesting that the final Pew poll gives Obama a 3% lead in the popular vote. Pew has been the best predictor in the past 2 presidential elections. Its final 2004 poll predicted 51 to 48 Bush, which was EXACTLY what it was, and in 2008 predicted 52 to 46 Obama; it was actually 53 to 46, or as close as you can come without hitting it exactly. The right hates Silver right now because he's saying Obama is almost certainly going to win. So they attack him, they attack his methods, his model, and his mother (just kidding about that one; though in my uncle's emails, it's close). They didn't do any of that in 2010, when he correctly predicted the Republican Congressional wave. The professional right pundits, you'll notice, are doing two things simultaneously now. 1). Saying Silver and the other aggregators are wrong, and Romney's going to win, just you watch; 2). Blaming the hurricane, or Chris Christie, or the media, or some other external force for Romney's coming loss. The former is what political operatives always do; the latter is what they do when they know they're going to lose.
Originally Posted By Labuda Thanks for the info I've gathered in here, y'all - I'm right now reading Nate Silver for the first time, and I'm liking his methods and manner of explanation! As for me, I'm with ecdc, I guess, who said this a couple of pages ago: I remain cautiously optimistic, but don't get out the confetti just yet, O supporters. ^^ That's me right now.
