t*b 发帖数: 850 | 1 很多人觉得诧异,为什莫不同poll出来的结果大不一样。
首先可以声明,poll是不会故意修改数据的,很多都是第三方代完成的,尤其是media
poll, 比如CNN, CBS的poll。
猫腻就出现在poll的对象怎末选。
poll绝不是随机抽取N个人调查。这N个人的组成才是poll的重中之重。一般来说,按照
一定比例选取一定名额的民主党,再选一定数量的共和党,最后是无党派人士。问题关
键在怎末定这个百分比。很多poll是按照2008年的投票人员组成来确定今年的抽样组成
的, 38%民主党, ~30%共和党,剩下的是无党派。这个比例准确性如何,仁者见仁,
智者见智。但是就算大部分民主党人也不会天真地认为今年民主党投票比例会有38%那
末高。
2004年,共和党投票比例也是高达38%,民主党31%-32%。可见这个比例才是关键。
如果你是亲奥巴马的,很有可能你会用民主党高比例的样品,反之亦成立。有些poll会
根据两党对这次大选的热情度,调整党派比例。
这个党派比例只是其中一个决定poll样品组成的一个重要因素。其他的因素也是需要考
虑在内的,比如黑人比例,白人比例,老墨比例。样品年龄段组成也要考虑在内。总的
来说今年的poll,如果按照2008组成进行,奥巴马会稍微领先,如果作了调整,罗姆尼
会领先。这就能解释为什莫所有的poll里面,在无党派人士中间罗姆尼都大幅领先奥巴
马5%-15%,为什莫有的poll却整体上落后于奥巴马。
有人说reputation, Media poll会看中这个?不会的。他们只关心力挺自己支持的候选
人。只有专门的poll机构,Gallup, Rasmussen才关心reputation.
有人说现在就是民主党人比共和党人,看看年轻人都是民主党的。这确实是这样。问题
是媒体不报,你也不知道的是很多人随着年龄的增长会变得越来越保守,这就是为什莫
大部分中老年人都是中间派或者共和党的。
具体到一个州的poll,那猫腻就更大了。Ohio为例,首先你要确定党派组成,然后再确
定地域组成,Cleavland 多少样品, Columbus多少抽样。这个比例那更是仁者见仁智
者见智了。
在一个左媒横行的年代社会,共和党候选人是得不到公平待遇的,从媒体报道的片面性
到poll的偏见。这就是我说的2008年没有出来投票的宗教人士,一些保守人士近年会出
来的,已经出来了,但他们不在poll里面。 | l**1 发帖数: 1875 | 2 Barack Obama to be re-elected President in 2012
Share on facebook Share on google_plusone
More Sharing Services
67.7%
CHANCE
Last prediction was: $6.77 / share
Today's Change: +$0.12 (+1.8%) | P*O 发帖数: 4324 | | g******y 发帖数: 1101 | 4 拿一年的统计数据毫无意义,根本没有任何普遍性。2004年是布什打宗教牌的一年,美
国保守教会倾巢出动,这才是当年共和党投票率高的根源。什么东西都是两面性的,这
些年美国年轻人信教比例每况愈下,和保守教会的所作所为不无关系。 | f*********g 发帖数: 1637 | 5 poll expert, which one you can trust, or what is your poll? | f*****g 发帖数: 15860 | 6 gallup历史polls和结果在这(1936-2008),还是很准的:
http://www.gallup.com/poll/9442/election-polls-accuracy-record-
media
【在 t*b 的大作中提到】 : 很多人觉得诧异,为什莫不同poll出来的结果大不一样。 : 首先可以声明,poll是不会故意修改数据的,很多都是第三方代完成的,尤其是media : poll, 比如CNN, CBS的poll。 : 猫腻就出现在poll的对象怎末选。 : poll绝不是随机抽取N个人调查。这N个人的组成才是poll的重中之重。一般来说,按照 : 一定比例选取一定名额的民主党,再选一定数量的共和党,最后是无党派人士。问题关 : 键在怎末定这个百分比。很多poll是按照2008年的投票人员组成来确定今年的抽样组成 : 的, 38%民主党, ~30%共和党,剩下的是无党派。这个比例准确性如何,仁者见仁, : 智者见智。但是就算大部分民主党人也不会天真地认为今年民主党投票比例会有38%那 : 末高。
| f*********g 发帖数: 1637 | 7 #6, According to Ramssaon (which all GOP deems a reliable pollster), if
sitting president job approval rate is or above 50%, he is very likely win
re-election. According to Gallup, Obama approval rating is 50% or more for
whole Oct. month now, and right now it's 51%. How do you reconcile two views
? | d****i 发帖数: 4354 | 8 你忘了Axelrod特别就approval rating关照过Gallup,后来Gallup的president
approval rating就高起来了。
http://www.breitbart.com/Big-Government/2012/09/06/gallup-doj-a
for
views
【在 f*********g 的大作中提到】 : #6, According to Ramssaon (which all GOP deems a reliable pollster), if : sitting president job approval rate is or above 50%, he is very likely win : re-election. According to Gallup, Obama approval rating is 50% or more for : whole Oct. month now, and right now it's 51%. How do you reconcile two views : ?
| f*********g 发帖数: 1637 | 9 You always like to quote the extremist site without thinking. If what they
said is correct, why Gallup wont raise Obama's poll? Instead Gallup poll is
the only consistent outlier at this moment. | d****i 发帖数: 4354 | 10 Gallup wants to keep their main poll intact, after all, that's their
livelihood. But they gave in on the approval rating as a compromise.
Also this is not a extremist site I believe:
http://www.washingtontimes.com/blog/watercooler/2012/sep/6/pick
I don't take this as fact, just a possibility.
is
【在 f*********g 的大作中提到】 : You always like to quote the extremist site without thinking. If what they : said is correct, why Gallup wont raise Obama's poll? Instead Gallup poll is : the only consistent outlier at this moment.
| | | f*********g 发帖数: 1637 | 11 "I don't take this as fact, just a possibility." okay your impression again,
bye. | d****i 发帖数: 4354 | 12 I'd like you to post something about who you support and why, see how much
fact you can put in without your own opinion.
It is obvious you don't like Romney or GOP, picking at others is always
easier.
again,
【在 f*********g 的大作中提到】 : "I don't take this as fact, just a possibility." okay your impression again, : bye.
| f******o 发帖数: 427 | 13 猫腻太多
media
【在 t*b 的大作中提到】 : 很多人觉得诧异,为什莫不同poll出来的结果大不一样。 : 首先可以声明,poll是不会故意修改数据的,很多都是第三方代完成的,尤其是media : poll, 比如CNN, CBS的poll。 : 猫腻就出现在poll的对象怎末选。 : poll绝不是随机抽取N个人调查。这N个人的组成才是poll的重中之重。一般来说,按照 : 一定比例选取一定名额的民主党,再选一定数量的共和党,最后是无党派人士。问题关 : 键在怎末定这个百分比。很多poll是按照2008年的投票人员组成来确定今年的抽样组成 : 的, 38%民主党, ~30%共和党,剩下的是无党派。这个比例准确性如何,仁者见仁, : 智者见智。但是就算大部分民主党人也不会天真地认为今年民主党投票比例会有38%那 : 末高。
| L*******9 发帖数: 936 | | t*b 发帖数: 850 | 15 you see, media bias is not about how they report events, it is about what
they choose to report, not to report.
Fox is conversative, they admit
MSNBC is libera, they admit
CNN, NBC, ABC, CBS are all leaning left, but they do not admit that.
When Romney's 46% revealed, all left media and Fox report it. Fox repeatedly
reported it in fact.
When Lybia attack was reavealed not due to the youtube film but it is a
terrorist attack, fox reported it. NO liberal media reported it. CNN
mentioned it a couple of times but keep implying Obama administration did
nothing wrong. NPR never said anything about that.
Until this past week, when Fox reported that CIA Annex forces want to help
the one-mile away embassy but denied, CNN, NPR all jumped in, started to
reporting that the fox report is wrong. NPR still did not report the attack
is a terrorist attack at all.
This is how the liberal media bias. I agree with you all, fox is very
conversative. but fox does not report all goo, bad things of both GOP and
DEM candidates. But CNN, NPR etc left media only report bad things of ROmney
, good things of Obama.
I think this is not bias anymore, it is actually crime, using media's
previlige try to alter the election result. | e****M 发帖数: 280 | 16 Registered voters的Polls都会报告民主党,共和党和无党派被调查的比例,所以有没
有bias是可以看出来的。当然调查公司可以撒谎,受众也可以选择不相信。
Likely voters的polls比较麻烦,因为不同的公司或者组织会用不同的模型和公式,而
且都是商业秘密,所以没法说。这些模型也不是简单的把以前投票比例拿过来用。poll
的时候会问其他关于选举的问题,比如投票地点,或者直接问你打不打算去投票,然后
通过这些问题的回答来判断是不是likely voter。
当然Gallup模型比其他公司的会好一点。不过这不是因为其他公司不关心自己声誉,主
要是模型难建,需要nb的人和很多年的修正。
媒体也不经常自己雇人去搞polls,更多的是报道专业调查公司的结果。
大部分媒体的是不是liberal很难说,左翼的人还觉得CNN偏右呢。但是fox确实是极右
,它想不承认也没办法。Fox的选择性报道和报道没有证实的消息是出了名的。当初共
和党初选的时候,Fox可是一点机会都没给Ron Paul,即使Ron Paul是最好的候选人。
如果一个人自己是极右,但是仍认为自己是偏右,那么中间立场的报道对来这个人来说
说就是偏左。同样,对一个自认为偏左的极左分子,中间立场的报道就是偏右。
美国的整体历史趋势确实是偏liberal。从独立战争到废奴到女性选举权到民权法案,
都是一步一步在向liberal迈进。其实整个人类历史都是倾向liberal的。偏不偏
liberal也和受教育的程度有关,受过高等教育的人偏liberal的多一些。现在这代中年
人受高等教育的比例偏低,所以偏conservative并不奇怪。偏liberal也不是坏事,要
不现在美国还有排华法案呢。 | t***h 发帖数: 5601 | 17 误差总是会有的, 大体上准确就行.
【在 f*****g 的大作中提到】 : gallup历史polls和结果在这(1936-2008),还是很准的: : http://www.gallup.com/poll/9442/election-polls-accuracy-record- : : media
| t*b 发帖数: 850 | 18 and you are wrong about America is leaning liberal.
overall the majority of people are still center right. If youa re true, how
a GOP candidate can win a national election? people lean left, media lean
left.
from what you say, it is clear you are a liberal. I am conversative, I admit
. but you just like other-than-MSBNC left meadia do not admit it. The reason
those media is leaning left, not because they want to be consistent with
American people. But it is simply because the ideology of the boss of thos
media.
CNN was in the middle 7-8 years ago, but since then CNN turns to the left
more and more. Do your own research, se how many conversative anchors have
been forced to leave CNN.
If you think CNN is stil in the middle, I really can not argue with you.
Right now, John King is th emost conversative person on CNN, he is probbaly
in the middle. The next is Erin Bunnet (Outfront). All others Cooper,
Blitzer, Candy are way on the left.
CNN's startegy is to exactly fool people like you. Only observe what is
happening on the surface. CNN does not report any negative things about any
DEM unless it is sexual scandal. Give me one example where CNN interviewed a
DEM and asked harsh questions (directly chanllenge the person). No you can
not find any.
Rather everytime CNN interviewed a GOP, compare the CNN's tone.
This is phony.
Fox says flat out they are converative. CNN has to hide. I even do not want
to mention MSNBC, which only said two things : 1) GOP is racist and 2) GOP
is criminals.
poll
【在 e****M 的大作中提到】 : Registered voters的Polls都会报告民主党,共和党和无党派被调查的比例,所以有没 : 有bias是可以看出来的。当然调查公司可以撒谎,受众也可以选择不相信。 : Likely voters的polls比较麻烦,因为不同的公司或者组织会用不同的模型和公式,而 : 且都是商业秘密,所以没法说。这些模型也不是简单的把以前投票比例拿过来用。poll : 的时候会问其他关于选举的问题,比如投票地点,或者直接问你打不打算去投票,然后 : 通过这些问题的回答来判断是不是likely voter。 : 当然Gallup模型比其他公司的会好一点。不过这不是因为其他公司不关心自己声誉,主 : 要是模型难建,需要nb的人和很多年的修正。 : 媒体也不经常自己雇人去搞polls,更多的是报道专业调查公司的结果。 : 大部分媒体的是不是liberal很难说,左翼的人还觉得CNN偏右呢。但是fox确实是极右
| t***h 发帖数: 5601 | 19 为什么不能随机挑选受访者呢? 例如在电话簿里随机挑选号码, 打电话询问.
为什么非要分党派分种族地来确定受访者?
确定地域组成,Cleavland 多少样品, Columbus多少抽样。这个比例那更是仁者见仁
智者见智了。
【在 t*b 的大作中提到】 : and you are wrong about America is leaning liberal. : overall the majority of people are still center right. If youa re true, how : a GOP candidate can win a national election? people lean left, media lean : left. : from what you say, it is clear you are a liberal. I am conversative, I admit : . but you just like other-than-MSBNC left meadia do not admit it. The reason : those media is leaning left, not because they want to be consistent with : American people. But it is simply because the ideology of the boss of thos : media. : CNN was in the middle 7-8 years ago, but since then CNN turns to the left
|
|