s*****o
发帖数: 1121 | 1 http://www.huffingtonpost.com/entry/donald-trump-debate-prep_us_57f6ea0ee4b0e655eab346d4
Marina Fang
Associate Politics Editor, The Huffington Post
Republican presidential nominee Donald Trump on Thursday held a rare town
hall event in New Hampshire that looked a lot like a trial run for Sunday’s
second presidential debate with Democrat Hillary Clinton.
But the softball questions selected from a Trump-friendly crowd offered
little practice for a nominee who desperately needs to overcome his wid... 阅读全帖 |
|
发帖数: 1 | 2 The mogul couldn’t even win over his supporters, the pollsters said: Only
37 percent of them said he is a role model, while more than half — 51
percent — said he is not. |
|
V****x
发帖数: 40 | 3 这个我觉得不算拉票，就是NBC/wsj电话poll，第一次被这些national pollster光顾，
所以决定花15看看他们搞什么。我本来就算independent，所以都实说，回答都是强烈
挺床反西。关于希拉里问了how concerned about Syria policy，没问email，
Wikileaks，所以这个poll也有倾向性。不过光问这些问题应该不会改变摇摆选民的立
场。
★ 发自iPhone App: ChineseWeb 13 |
|
T**********e
发帖数: 29576 | 4 当然 比较隐蔽，但确定把老头tape灌你耳朵里。2000年初选布什对麦肯在南卡push
poll更直接，问一堆问题然后问有没听说麦肯养私生子小黑孩，等弄明白小黑孩是领养
的SC都选完了
: 这个我觉得不算拉票，就是NBC/wsj电话poll，第一次被这些national
pollster
光顾，
: 所以决定花15看看他们搞什么。我本来就算independent，所以都实说，
回答都
是强烈
: 挺床反西。关于希拉里问了how concerned about Syria policy，没问
email，
: Wikileaks，所以这个poll也有倾向性。不过光问这些问题应该不会改变
摇摆选
民的立
: 场。
: ★ 发自iPhone App: ChineseWeb 13
|
|
i**********k
发帖数: 5274 | 5 Friday, October 14, 2016
At the close of a week that began with him trailing by seven points, Donald
Trump still holds a slight lead over Hillary Clinton in today’s White House
Watch survey despite a flurry of news reports alleging a history of sexual
harassment on his part.
The latest Rasmussen Reports national telephone and online survey of Likely
U.S. Voters shows Trump with 43% support to Clinton’s 41%. That’s
unchanged from yesterday. Libertarian candidate Gary Johnson picks up six
percent ... 阅读全帖 |
|
g******t
发帖数: 18158 | 6 打烂老婊砸走狗的碧莲
Criticism of presidential forecasts[edit]
In a series of posts in 2011 and 2012, FiveThirtyEight criticized the
forecasting methods that relied on macro-economic modeling of the electoral
outcomes.[538 80][538 81] According to Silver, models based primarily on the
macro-level performance of the economy (such as unemployment, inflation,
and the performance of the stock market), presidential approval ratings (
when an incumbent is running for re-election), and the ideological
positionin... 阅读全帖 |
|
g******t
发帖数: 18158 | 7 打烂老婊砸走狗的碧莲
Criticism of presidential forecasts[edit]
In a series of posts in 2011 and 2012, FiveThirtyEight criticized the
forecasting methods that relied on macro-economic modeling of the electoral
outcomes.[538 80][538 81] According to Silver, models based primarily on the
macro-level performance of the economy (such as unemployment, inflation,
and the performance of the stock market), presidential approval ratings (
when an incumbent is running for re-election), and the ideological
positionin... 阅读全帖 |
|
g******t
发帖数: 18158 | 8 打烂老婊砸走狗的碧莲
Criticism of presidential forecasts[edit]
In a series of posts in 2011 and 2012, FiveThirtyEight criticized the
forecasting methods that relied on macro-economic modeling of the electoral
outcomes.[538 80][538 81] According to Silver, models based primarily on the
macro-level performance of the economy (such as unemployment, inflation,
and the performance of the stock market), presidential approval ratings (
when an incumbent is running for re-election), and the ideological
positionin... 阅读全帖 |
|
s***5
发帖数: 453 | 9 http://www.latimes.com/politics/la-na-pol-polling-differences-20160809-snap-story.html
A couple of other aspects of the Daybreak poll make it differ from most
other surveys:
Typically, polls ask people which candidate they favor or lean toward. Those
who say they don’t know or are undecided don’t get factored into
calculations of candidate support.
The Daybreak poll, by contrast, asks voters, using a 0-to-100 scale, to rate
their chances of voting for Clinton, for Trump or for some other candida... 阅读全帖 |
|
y***r
发帖数: 16594 | 10 Never before in the 20-year history of Latina has the magazine endorsed a
candidate—until today.
But as the magazine noted, this is not an ordinary election. So, Latina
broke tradition, joining forces with leaders in media and business, artists,
and influencers to send a strong message—that we stand together to endorse
Hillary Clinton for president.
The roster includes powerhouse Latinos and organizations: Dolores Huerta,
Eva Longoria, John Leguizamo, the Hispanic Chamber of Commerce, Planned
Pa... 阅读全帖 |
|
z*****8
发帖数: 1051 | 11 这些pollsters慢慢会放真实数据出来的，要不以后谁买他们的货，都是找个台阶下。 |
|
l***o
发帖数: 7937 | 12 pollsters 要养活自己啊。做民调前，先联系好稀婆阵营，问要不要给点赞助啥的，你
懂的 |
|
l****e
发帖数: 609 | 13 谢谢您的知识。受教了。我是外行，但如果群体不符合选民的分布,可以用上权重。
方舟子说
“所以，要对全美国做民意调查，随机抽查300人也是可以的，川普、川粉的
嘲笑只不过暴露了他们对统计学的无知。不过，抽查300人的误差有些大，全国
性民意调查的误差一般控制在3%左右，因此需要的样本大小通常是1000人左右。
CNN/ORC调查了521人，误差是4.5%，误差稍大，但是因为调查的结果比率悬殊
（62%:27%），这个误差并不会改变希拉里大胜的结论。CNN/ORC调查采用的是简
单随机抽样的方式，这样抽得的随机样本不一定能很好地符合人群分布特征，例
如其中认同民主党的人数偏高。所以有时候为了获得能代表人群分布的样本，会
采取分层抽样的方式，先根据人群特征设定各个群体的比例，然后再在各个群体
随机抽样。”
CNN用的是"random-digit dialling,"，见下面的英文
http://edition.cnn.com/ALLPOLITICS/1996/polls/FAQ/polling.FAQ.shtml#howinterview
方舟子还说
“因为是抽样调查，即使精心地设计，也不能... 阅读全帖 |
|
z*****8
发帖数: 1051 | 14 setting up the samples in biased fashion is a confession of the intention of
the pollster. There are many tricks you can play throughout the process to
skew the result. Everyone is entitled to their own opinion. But to convince
others, one needs to know about the subject in depth, not just echo-
chambering the so-called legit main stream media with full "authority". |
|
l****e
发帖数: 609 | 15 自己读读英文吧，Simple random sampling也是一种Probability Sampling Methods，
我不是搞这行的，没有发言权，但我相信方舟子比你和这里绝大多数川粉强，人家说得
有理有据。
说了这么久，我怎么听不出你的高明之处，你能否把一个复杂的问题简单化了讲？你既
然单挑我，说随机抽查不科学，你总得说出过原由吧。
现在这个回帖，你不再是在说随机抽查的不科学性，而是在说CNN的抽样的电话号码不
是随机的，是在质疑CNN的credibility吗?
Probability Sampling Methods:
1.Simple random sampling
2.Stratified sampling
3.Cluster sampling
4.Multistage sampling
5.Systematic random sampling
另外www.aapor.org有关于Random-Digit Dialing (RDD)的介绍
http://www.aapor.org/Education-Resources/Election-Polling-Resourc... 阅读全帖 |
|
发帖数: 1 | 16 樓主賣了半天的關仔, 好像stratified sampling是甚麼很神秘的東西, 何必呢? 有話
直說不是很好?
美國有三億人口, 還是四億, 為了達到 3% 的誤差, 可能要打電話給三千人. 但是這數
目太大, pollster 不容易做到, 所以統計學說, 可以把這三四億人口 homogeneously
地分成幾塊(稱作 strata), 例如分成一百塊, 每塊人口是三四百萬. 這樣子,只要在選
定的一塊中隨機抽樣三十人, 就可以達到 3% 的誤差. 這樣的民調比較容易做到.
注意關鍵詞是 homogeneous. 意思是說, 每一塊里的人口組成在各方面都一樣, 性別,
黨派, 收入, 年齡....等等. 所以分 strata 不是那麼容易, 你不能隨便拿一個電話區
域號碼代表權美國, 肯定有錯.
樓主一直賣關子, 不肯/不敢明說, 但我猜他的意思是說 CNN 在分 strat 的時候就動
了手腳, 裡面的人群不是同質性的.
這一樓裡面幾個插科打渾的角色如 fishingarden 上了一堆帖子, 但完全講不出內容,
就是罵街, 另外還有幾個如 huhu 根本就是近來罵街的. ... 阅读全帖 |
|
g********2
发帖数: 6571 | 17 这话说得很有pollster的风度了。
80% |
|
e********t
发帖数: 397 | 18 http://wpo.st/a4y62
When the facts don’t matter,
how can democracy survive?
Americans — or, at least, a particular subset of Americans — have had
enough of experts, facts, math, data. They distrust them all.
This rising cynicism, sown recklessly by opportunistic politicians, will
not only make it increasingly difficult for policymakers to make good
choices and govern peacefully; it could also become a significant
economic challenge.
The latest evidence of this anti-evidence trend comes from... 阅读全帖 |
|
e********t
发帖数: 397 | 19 http://wpo.st/a4y62
When the facts don’t matter,
how can democracy survive?
Americans — or, at least, a particular subset of Americans — have had
enough of experts, facts, math, data. They distrust them all.
This rising cynicism, sown recklessly by opportunistic politicians, will
not only make it increasingly difficult for policymakers to make good
choices and govern peacefully; it could also become a significant
economic challenge.
The latest evidence of this anti-evidence trend comes from... 阅读全帖 |
|
|
h*******n
发帖数: 8906 | 21 只看poll 别的一律不看 也很幼稚园的行为 人专业pollster是要考虑enthusiasm
momentum 修正参数呢
: 说自己领先就是好民调，其他都是rigged。丫就一幼儿园水平。
|
|
h*******n
发帖数: 8906 | 22 你的假设有意义么 pollster很大程度也是基于假设了
: What if his prediction is correct?
: Polls
|
|
h*******n
发帖数: 8906 | 23 你的假设有意义么 pollster很大程度也是基于假设了
: What if his prediction is correct?
: Polls
|
|
f********8
发帖数: 5601 | 24 Pollsters: We cannot poll this thing. It's a movement. |
|
d******y
发帖数: 2787 | 25 “我三八”，类似pollster中的绿茶婊
★ 发自iPhone App: ChineseWeb 13 |
|
|
发帖数: 1 | 27 No shock for deplorables! HaHa |
|
d******y
发帖数: 2787 | 28 We are ready for whatever shocks! Haha
★ 发自iPhone App: ChineseWeb 13 |
|
h*******n
发帖数: 8906 | 29 Per foxnews pollster Frank Luntz. Again it is absentee ballot return count
according to party line. Not the final vote result.
Early vote returns in NORTH CAROLINA so far
Democrats: 17,543 in 2016 (17,331 in 2012)
Republicans: 19,533 in 2016 (34,103 in 2012)
Early vote returns in FLORIDA so far
Democrats: 305,700 in 2016 (111,471 in 2008)
Republicans: 314,301 in 2016 (176,855 in 2008)
Early vote returns in IOWA so far
Democrats: 99,345 in 2016 (132,188 in 2012)
Republicans: 61,065 in 2016 (78... 阅读全帖 |
|
h*******n
发帖数: 8906 | 30 Per foxnews pollster Frank Luntz. Again it is absentee ballot return count
according to party line. Not the final vote result.
Early vote returns in NORTH CAROLINA so far
Democrats: 17,543 in 2016 (17,331 in 2012)
Republicans: 19,533 in 2016 (34,103 in 2012)
Early vote returns in FLORIDA so far
Democrats: 305,700 in 2016 (111,471 in 2008)
Republicans: 314,301 in 2016 (176,855 in 2008)
Early vote returns in IOWA so far
Democrats: 99,345 in 2016 (132,188 in 2012)
Republicans: 61,065 in 2016 (78... 阅读全帖 |
|
t*c
发帖数: 6929 | 31 我根本不相信poll，其实大部分人也只是参考。
poll的科学性在于取样的公平性，如果一个机构
有明显的政治偏向性，那么这些poll根本就无效。
08年，如果看polls，希拉里早就获得民主党提名，
不会输给奥巴马。2016年民主党密歇根初选的时候，ABC，
CNN的poll都给希拉里15%左右的领先，Sanders却以
2%的优势击败了希拉里。
BBC在报道英国脱欧投票当天，以90%的概率英国会留欧，
而且镜头都对准那些支持留欧的人群，以至于全世界都在
想英国留欧会以大比数通过，直到纽卡斯尔开票，这个一直
民调留欧派领先10%以上的英格兰城市，居然以五五开，尴尬的BBC主持人不得不说，
maybe
today is not the day we have expected.
所以川粉希粉都不要争，那没用，无论如何一定要去投票，投完票，大家才会知道真相
！不去投票，你就上了有些pollster 的当了! |
|
发帖数: 1 | 32 Donald Trump’s campaign manager admits: ‘We are behind’
MIAMI - Donald Trump's campaign manager Kellyanne Conway, a longtime
Republican pollster, admitted Sunday that her candidate is currently losing
to Democrat Hillary Clinton.
"We are behind," Conway said on NBC's "Meet the Press."
Clinton, she said, "
has some advantages, like $66 million in ad buys just in the month of
September, thereby doubling her ad buys from August. Now, most of those ads
are negative against Donald T... 阅读全帖 |
|
发帖数: 1 | 33 How to Legally Vote More Than Once
Posted October 23rd, 2016 @ 8:04pm in #Trump #Clinton
If you find it annoying that you only have one legal vote, here’s how you
can get a few more. It’s called persuasion.
You can multiply the power of your opinion by convincing people on the other
side to stay home on election day. Every vote you suppress on the other
side is like an extra vote for you. And there’s no limit to how many you
can have!
Persuasion doesn’t work every time. But you might enjoy exper... 阅读全帖 |
|
c*******o
发帖数: 8869 | 34 有名声的POLLSTER还是重视track record的。 gallup在2012错的离谱，所以USA TODAY
马上就fire了他们。 |
|
|
c*******o
发帖数: 8869 | 36 "And in January, Gallup and USA Today ended their 20-year polling
partnership. While both parties described the breakup as amicable, the
pollster’s misfire on the 2012 election loomed large in the background" |
|
t******r
发帖数: 8600 | 37 90% pollsters are democRATs。说这些人有节操，你是要笑死人吗？！
: 做poll的人没那么没节操，要改方法也只会在大选后改，尤其这种每天收集数据
的
: tracking poll。
|
|
q***0
发帖数: 225 | 38 看看历史，rally人数超多但输掉大选的比比皆是，输的一方怀疑poll的也很多。2012
就是一个例子，历史上 Mondale，杜卡基斯等民主党的例子更明显。看看 Mondale vs
Reagan 时，Mondale怎么说：
"There's something going on in this country and the pollsters aren't getting
it,"
"Nobody who's been with me for the last few days and has seen these crowds,
seen their response, seen their enthusiasm, seen the intensity of their
response, and how they respond to these issues ... can help but believe
there is something happening in this country."
太相似了，只是 Mondale 更 eloquent |
|
l***o
发帖数: 7937 | 39 今年媒体、政党、pollster大联姻，丢了可信度。 |
|
b*******e
发帖数: 724 | 40 rep已经way ahead ofpace than 2012。关键的是从眼下voting的情况看，今年poll误
差非常大，trump的很多voter是第一次vote，或者很多年没有vote了，这些人的热情非
常高，几乎100%的likely voter但是pollster从来没有调查这些人，相反妖婆的voter
大多是OB的voter，所以很容易oversample这些人，这就是为什么今年polls非常不准。
看看TX的earlyvoting，2012年到现在才60几万人投票，现在已经过1米了，这些多出来
的voter大多是投trump的。现在trump还需要再赢一个主要的toss upstate。眼下有可
能的是：PA, WI, MI, CO, NH MN 我觉得随着妖婆的email丑闻继续发酵，赢至少一个
的可能性超过70%，有可能trump将这几个州一锅端。所以不出以为，2016 election已
经结束了 |
|
b*******e
发帖数: 724 | 41 rep已经way ahead ofpace than 2012。关键的是从眼下voting的情况看，今年poll误
差非常大，trump的很多voter是第一次vote，或者很多年没有vote了，这些人的热情非
常高，几乎100%的likely voter但是pollster从来没有调查这些人，相反妖婆的voter
大多是OB的voter，所以很容易oversample这些人，这就是为什么今年polls非常不准。
看看TX的earlyvoting，2012年到现在才60几万人投票，现在已经过1米了，这些多出来
的voter大多是投trump的。现在trump还需要再赢一个主要的toss upstate。眼下有可
能的是：PA, WI, MI, CO, NH MN 我觉得随着妖婆的email丑闻继续发酵，赢至少一个
的可能性超过70%，有可能trump将这几个州一锅端。所以不出以为，2016 election已
经结束了 |
|
d******y
发帖数: 2787 | 42 亮点是pollster
★ 发自iPhone App: ChineseWeb 13 |
|
d********f
发帖数: 43471 | 43 这是和党的御用pollster,ｃｏ主要是年轻人多，所以turn out比ｐｏｌｌ重要 |
|
t*c
发帖数: 6929 | 44 其实媒体手里都有各地pollsters提供的raw data，
比如A P，就看他们会不会公布，何时公布了。
: 看看英国对公投干的事 我对美国民主的有效性有点怀疑
|
|
T**********e
发帖数: 29576 | 45 几个根据计算选举人票的模型，老太胜选机会：
HuffPost Pollster and the Princeton Election Consortium： close to 100
percent。 Daily Kos Elections’ model：91 percent.
NYT Upshot model：84 percent |
|
q***0
发帖数: 225 | 46 这个 Harper polling 也是共和党的pollster，和那个Remington 一样。这两天共和党
真是推出不少polls，呵呵 |
|
q***0
发帖数: 225 | 47 这个 Harper polling 也是共和党的pollster，和那个Remington 一样。这两天共和党
真是推出不少polls，呵呵 |
|
g********2
发帖数: 6571 | 48 装13！
楼主又不是pollster，发表的又不是poll结果，只不过是人家的个人感受，你装什么13？
你这么懂历史，你来说说你poll的结果？
不读历史真可怕。去看看当初盖洛普是咋发家的。你的顾客龙虾党有几个？收入是均
总统。 |
|
g********2
发帖数: 6571 | 49 装13！
楼主又不是pollster，发表的又不是poll结果，只不过是人家的个人感受，你装什么13？
你这么懂历史，你来说说你poll的结果？
不读历史真可怕。去看看当初盖洛普是咋发家的。你的顾客龙虾党有几个？收入是均
总统。 |
|
|
