Polls, polls, polls, ARE THEY ACCURATE?!

[hazek]

I wrote this in another thread but I thought it deserved it's own. There's just too many of you out there who seem to not want to learn from just 4 years ago. And to those of you I say this:

"Those Who Forget the Lessons of History Are Doomed To Repeat It" - John A. N. Lee

In 2007 we also thought the polls couldn't possibly be the picture of reality, and yet they turned out to be remarkably accurate come election day.

Of course the polls changed throughout the process but in the end just a week before the election they were very accurate.

To get a good picture of this please go to http://en.wikipedia.org/wiki/Statewide_opinion_polling_for_the_Republican_Party _presidential_primaries,_2008

EXAMPLE: 2007/2008 IOWA POLLS::
--------------------------

Caucus Results
Sampling Size: 118,696
January 3, 2008 Mike Huckabee 34.4%, Mitt Romney 25.2%, Fred Thompson 13.4%, John McCain 13.1%, Ron Paul 10%, Rudy Giuliani 3.5%, Duncan Hunter .4%

Poll Source Date Highlights

Insider Advantage January 1, 2008 Mike Huckabee 30%, Mitt Romney 24.1%, Fred Thompson 11.3%, John McCain 10.9%, Ron Paul 7.1%, Rudy Giuliani 4.8%, Duncan Hunter 1.3%, Undecided 10.5%
American Research Group
Sample Size: 600
Margin of Error: ±4%
December 31, 2007 - January 2, 2007 Mike Huckabee 29%, Mitt Romney 24%, Fred Thompson 13%, John McCain 11%, Rudy Giuliani 8%, Ron Paul 6%, Duncan Hunter 4%, Alan Keyes 1%, Undecided 4%
Zogby International
Sample Size: 914
Margin of Error: ±3.3%
December 30, 2007 - January 2, 2008 Mike Huckabee 31%, Mitt Romney 25%, Fred Thompson 11%, John McCain 10%, Ron Paul 10%, Rudy Giuliani 6%, Duncan Hunter 1%, Undecided 6%
Zogby International
Sample Size: 882
Margin of Error: ±3.3%
December 29, 2007 - January 1, 2008 Mike Huckabee 28%, Mitt Romney 26%, John McCain 12%, Fred Thompson 12%, Ron Paul 9%, Rudy Giuliani 7%, Duncan Hunter 1%, Undecided 6%
Zogby International
Sample Size: 903
Margin of Error: ±3.3%
December 28–31, 2007 Mike Huckabee 29%, Mitt Romney 25%, John McCain 12%, Fred Thompson 10%, Rudy Giuliani 8%, Ron Paul 7%, Duncan Hunter 1%, Undecided 6%
Strategic Vision (note)
Sample Size: 600
Margin of Error: ±4.5%
December 28–30, 2007 Mitt Romney 30%, Mike Huckabee 28%, John McCain 16%, Fred Thompson 13%, Ron Paul 4%, Rudy Giuliani 4%, Duncan Hunter 1%, Undecided 4%
Des Moines Register
Sample Size: 800
Margin of Error: ±3.5%
December 27–30, 2007 Mike Huckabee 32%, Mitt Romney 26%, John McCain 13%, Ron Paul 9%, Fred Thompson 9%, Rudy Giuliani 5%, Alan Keyes 1%, Duncan Hunter 1%, Undecided 4%
Zogby International
Sample Size: 876
Margin of Error: ±3.3%
December 27–30, 2007 Mike Huckabee 29%, Mitt Romney 27%, John McCain 13%, Fred Thompson 8%, Rudy Giuliani 7%, Ron Paul 7%, Duncan Hunter 1%, Undecided 6%
CNN/Opinion Research Corp.
Sample Size: 373
Margin of Error: ±5%
December 26–30, 2007 Mitt Romney 31%, Mike Huckabee 28%, Fred Thompson 13%, John McCain 10%, Rudy Giuliani 8%, Ron Paul 8%, Duncan Hunter <0.5%
Zogby International
Sample Size: 867
Margin of Error: ±3.4%
December 26–29, 2007 Mike Huckabee 29%, Mitt Romney 28%, John McCain 11%, Fred Thompson 8%, Rudy Giuliani 8%, Ron Paul 8%, Duncan Hunter 1%, Undecided 6%
MSNBC/Mason-Dixon
Sample Size: 400
Margin of Error: ±5%
December 26–28, 2007 Mitt Romney 27%, Mike Huckabee 23%, Fred Thompson 14%, John McCain 13%, Rudy Giuliani 5%, Ron Paul 5%, Duncan Hunter 1%, Undecided 12%
American Research Group
Sample Size: 600
Margin of Error: ±4%
December 26–28, 2007 Mitt Romney 32%, Mike Huckabee 23%, John McCain 11%, Fred Thompson 7%, Rudy Giuliani 6%, Ron Paul 6%, Duncan Hunter 2%, Undecided 11%
Research 2000/Sioux City Journal
Sample Size: 500
Margin of Error: ±4.5%
December 26–27, 2007 Mike Huckabee 34%, Mitt Romney 27%, Fred Thompson 11%, John McCain 8%, Rudy Giuliani 8%, Ron Paul 8%, Duncan Hunter 1%, Undecided 3%
Strategic Vision (R)
Sampling Size: 600LV
Margin of Error: ± 4.5%
December 26–27, 2007 Mike Huckabee 29%, Mitt Romney 27%, Fred Thompson 15%, John McCain 14%, Rudy Giuliani 4%, Ron Paul 4%, Duncan Hunter 1%, Undecided 6%
LA Times/Bloomberg December 20–23, 26, 2007 Mike Huckabee 37%, Mitt Romney 23%, John McCain 11%, Fred Thompson 11%, Rudy Giuliani 6%, Ron Paul 2%, Duncan Hunter 1%, Undecided 9%
American Research Group
Sampling Size: 600
Margin of Error: ± 4%
December 20–23, 2007 Mike Huckabee 23%, Mitt Romney 21%, John McCain 17%, Rudy Giuliani 14%, Ron Paul 10%, Fred Thompson 3%, Duncan Hunter 2%, Alan Keyes 2%, Undecided 8%
American Research Group
Sampling Size: 600
Margin of Error: ± 4%
December 16–19, 2007 Mike Huckabee 28%, John McCain 20%, Mitt Romney 17%, Rudy Giuliani 13%, Fred Thompson 5%, Ron Paul 4%, Tom Tancredo 1%, Alan Keyes 1%, Duncan Hunter -%, Undecided 11%
Strategic Vision (R)
Sampling Size: 600LV
Margin of Error: ± 4.5%
December 16–18, 2007 Mike Huckabee 31%, Mitt Romney 25%, Fred Thompson 16%, John McCain 8%, Rudy Giuliani 6%, Ron Paul 5%, Tom Tancredo 2%, Duncan Hunter 1%, Undecided 6%
Rasmussen Reports
Sampling Size: 496
Margin of Error: ± 4%
December 17, 2007 Mike Huckabee 28%, Mitt Romney 27%, John McCain 14%, Rudy Giuliani 8%, Fred Thompson 8%, Ron Paul 6%, Tom Tancredo 2%, Duncan Hunter 1%
Strategic Vision (R)
Sampling Size: 600LV
Margin of Error: ± 4.5%
December 8–10, 2007 Mike Huckabee 30%, Mitt Romney 25%, Fred Thompson 13%, Rudy Giuliani 10%, John McCain 5%, Ron Paul 4%, Tom Tancredo 3%, Duncan Hunter 1%, Undecided 9%
Rasmussen Reports
Sampling Size: 789
December 10, 2007 Mike Huckabee 39%, Mitt Romney 23%, Fred Thompson 8%, Rudy Giuliani 8%, John McCain 6%, Ron Paul 5%, Tom Tancredo 3%, Duncan Hunter 1%
Newsweek
Sampling Size: 540
Margin of Error: ± 3%
December 5–6, 2007 Mike Huckabee 39%, Mitt Romney 17%, Fred Thompson 10%, Rudy Giuliani 9%, Ron Paul 8%, John McCain 6%, Tom Tancredo 2%, Duncan Hunter 0%, Other 1%, Undecided 8%
Mason-Dixon
Sampling Size: 400
Margin of Error: ± 5%
December 3–6, 2007 Mike Huckabee 32%, Mitt Romney 20%, Fred Thompson 11%, John McCain 7%, Rudy Giuliani 5%, Ron Paul 2%, Duncan Hunter 1%, Alan Keyes 1%, Undecided 19%
Strategic Vision (note)
Sampling Size: 600
Margin of Error: ± 4.5%
November 30 - December 2, 2007 Mike Huckabee 27%, Mitt Romney 24%, Rudy Giuliani 13%, Fred Thompson 11%, John McCain 6%, Ron Paul 5%, Tom Tancredo 2%, Duncan Hunter 1%, Undecided 11%
American Research Group
Sampling Size: 600
Margin of Error: ± 4%
November 26–29, 2007 Mitt Romney 28%, Mike Huckabee 27%, Fred Thompson 14%, Rudy Giuliani 9%, John McCain 9%, Ron Paul 3%, Tom Tancredo 1%, Duncan Hunter -%, Alan Keyes -%, Undecided 9%
Des Moines Register November 25–28, 2007 Mike Huckabee 29%, Mitt Romney 24%, Rudy Giuliani 13%, Fred Thompson 9%, John McCain 7%, Ron Paul 7%, Tom Tancredo 6%, Duncan Hunter 1%, John Cox -%, Alan Keyes -%, Undecided 4%
Rasmussen November 26–27, 2007 Mike Huckabee 28%, Mitt Romney 25%, Rudy Giuliani 12%, Fred Thompson 11%, Ron Paul 5%, John McCain 4%
Strategic Vision (note)
Sampling Size: 600
Margin of Error: ±4.5%
November 23–25, 2007 Mitt Romney 26%, Mike Huckabee 24%, Rudy Giuliani 14%, Fred Thompson 10%, John McCain 7%, Ron Paul 5%, Tom Tancredo 3%, Duncan Hunter 1%, Undecided 10%
Pew Research Center
Sampling Size: 264
Margin of Error: ±7%
November 7–25, 2007 Mitt Romney 25%, Mike Huckabee 24%, Rudy Giuliani 14%, Fred Thompson 12%, John McCain 5%, Ron Paul 5%, Tom Tancredo 3%, Duncan Hunter 1%, Other 10%
ABC News/Wash Post November 14–18, 2007 Mitt Romney 28%, Mike Huckabee 24%, Fred Thompson 15%, Rudy Giuliani 13%, John McCain 6%, Ron Paul 6%, Tom Tancredo 2%, Duncan Hunter 1%, Undecided 4%
KCCI Des Moines
Sampling Size: 600
Margin of Error: ± 4%
November 12–14, 2007 Mitt Romney 27%, Mike Huckabee 18%, Rudy Giuliani 16%, Fred Thompson 10%, John McCain 6%, Ron Paul 5%, Tom Tancredo 2%, Duncan Hunter 1%, Undecided 15%
American Research Group
Sampling Size: 600
Margin of Error: ± 4%
November 10–14, 2007 Mitt Romney 26%, Mike Huckabee 24%, Rudy Giuliani 11%, Fred Thompson 11%, John McCain 10%, Ron Paul 3%, Tom Tancredo 1%, Duncan Hunter 1%, Keyes -, Undecided 13%
Rasmussen Reports
Sampling Size: 825
Margin of Error: ± 3.5%
November 9–12, 2007 Mitt Romney 29%, Mike Huckabee 16%, Rudy Giuliani 15%, Fred Thompson 14%, John McCain 6%, Ron Paul 4%, Tom Tancredo 4%, Duncan Hunter 2%, Undecided 9%
Strategic Vision (note)
Sampling Size: 600
Margin of Error: ± 4%
November 9–12, 2007 Mitt Romney 30%, Mike Huckabee 19%, Rudy Giuliani 12%, Fred Thompson 11%, John McCain 7%, Ron Paul 5%, Tom Tancredo 2%, Duncan Hunter 1%, Undecided 13%
CBS News/New York Times
Sampling Size: 1273
Margin of Error: +/- 5%
November 2–11, 2007 Mitt Romney 27%, Mike Huckabee 21%, Rudy Giuliani 15%, Fred Thompson 9%, Ron Paul 4%, John McCain 4%, Tom Tancredo 3%, Duncan Hunter 2%, Undecided 14%
Zogby
Sampling Size: 410
Margin of Error: +/- 5.0%
November 6–7, 2007 Mitt Romney 31%, Mike Huckabee 15%, Rudy Giuliani 11%, Fred Thompson 10%, John McCain 8%, Ron Paul 4%, Tom Tancredo 3%, Duncan Hunter 1%, Other 1%, Not Sure 16%
American Research Group October 26–29, 2007 Mitt Romney 27%, Mike Huckabee 19%, Rudy Giuliani 16%, John McCain 14%, Fred Thompson 8%, Tom Tancredo 2%, Ron Paul 1%, Duncan Hunter -%, Alan Keyes -%, Undecided 13%
University of Iowa Hawkeye Poll
Margin of Error: +/- 5.8%
Sampling Size: 285
October 17–24, 2007 Mitt Romney 36.2%, Rudy Giuliani 13.1%, Mike Huckabee 12.8%, Fred Thompson 11.4%, John McCain 6.0%, Tom Tancredo 2.2%, Others 3.5%, Undecided 14.9%
Strategic Vision (note) October 12–14, 2007 Mitt Romney 28%, Rudy Giuliani 13%, Mike Huckabee 12%, Fred Thompson 10%, John McCain 5%, Sam Brownback 4%, Ron Paul 4%, Tom Tancredo 2%, Duncan Hunter 1%, Undecided 22%
Rasmussen Reports October 10 & 14, 2007 Mitt Romney 25%, Fred Thompson 19%, Mike Huckabee 18%, Rudy Giuliani 13%, John McCain 6%, Sam Brownback 3%, Ron Paul 2%, Tom Tancredo 2%, Duncan Hunter 1%, Undecided 11%
InsiderAdvantage October 2–3, 2007 Mitt Romney 24%, Rudy Giuliani 16%, Fred Thompson 13%, Mike Huckabee 13%, John McCain 10%, Sam Brownback 5%, Ron Paul 4%, Duncan Hunter 2%, Undecided 13%
Des Moines Register October 1–3, 2007 Mitt Romney 29%, Fred Thompson 18%, Mike Huckabee 12%, Rudy Giuliani 11%, John McCain 7%, Tom Tancredo 5%, Ron Paul 4%, Sam Brownback 2%, Alan Keyes 2%, Duncan Hunter 1%, John Cox -, Undecided 9%
American Research Group September 26–29, 2007 Mitt Romney 22%, Rudy Giuliani 21%, Fred Thompson 16%, John McCain 11%, Newt Gingrich 5%, Mike Huckabee 4%, Ron Paul 2%, Duncan Hunter 2%, Sam Brownback 2%, Alan Keyes 1%, Tom Tancredo 1%, Undecided 13%
Newsweek (All Republican voters) September 26–27, 2007 Mitt Romney 25%, Fred Thompson 16%, Rudy Giuliani 15%, John McCain 7%, Mike Huckabee 6%, Tom Tancredo 3%, Sam Brownback 3%, Ron Paul 2%, Duncan Hunter 1%, Undecided 21%
Newsweek (Likely caucus-goers) September 26–27, 2007 Mitt Romney 24%, Fred Thompson 16%, Rudy Giuliani 13%, Mike Huckabee 12%, John McCain 9%, Ron Paul 5%, Tom Tancredo 3%, Sam Brownback 2%, Duncan Hunter 1%, Undecided 15%
Strategic Vision (note) September 21–23, 2007 Mitt Romney 30%, Rudy Giuliani 17%, Fred Thompson 13%, Mike Huckabee 8%, John McCain 6%, Newt Gingrich 5%, Ron Paul 3%, Tom Tancredo 3%, Sam Brownback 2%, Duncan Hunter 1%, Undecided 13%
Los Angeles Times/Bloomberg Poll September 6–10, 2007 Mitt Romney 28%, Rudy Giuliani 16%, Fred Thompson 16%, Mike Huckabee 8%, John McCain 7%, Tom Tancredo 3%, Ron Paul 2%, Sam Brownback 2%, Duncan Hunter 1%, Undecided 15%
American Research Group August 26–29, 2007 Mitt Romney 27%, Rudy Giuliani 17%, Mike Huckabee 14%, Fred Thompson 13%, Newt Gingrich 7%, John McCain 5%, Hunter 1%, Paul 1%, Brownback -, Tancredo -, Undecided 15%
McLaughlin & Associates August 20–21, 2007 Mitt Romney 35%, Rudy Giuliani 12%, Fred Thompson 11%, Mike Huckabee 11%, Tom Tancredo 9%, John McCain 7%, Sam Brownback 2%, Ron Paul 1%, Duncan Hunter 1%, Firm Undecided 10%
Strategic Vision (note)
(Likely Caucus Goers)
Margin of Error: +/- 4%
Sampling Size: 600
August 17–19, 2007 Mitt Romney 31%, Fred Thompson 15%, Rudy Giuliani 13%, John McCain 8%, Mike Huckabee 8%, Newt Gingrich 3%, Ron Paul 3%, Tom Tancredo 3%, Sam Brownback 2%, Duncan Hunter 1%, Undecided 13%
Newsmax/Zogby
(Likely Caucus Goers)
Margin of Error: +/- 4.5%
Sampling Size: 487
August 17–18, 2007 Mitt Romney 33%, Rudy Giuliani 14%, Fred Thompson 12%, Mike Huckabee 8%, John McCain 6%, Sam Brownback 4%, Tom Tancredo 3%, Ron Paul 3%, Duncan Hunter 1%, Someone else 2%, Not sure 14%

[Cowlesy]

I imagine the polls are reasonably accurate and like 2007, very disappointing. But, let's press onward and do the best we can.

[RonPaulCult]

Ok but one national poll says 5% and one says 13%. Which one should we believe?

[afwjam]

We are at about 8% nationally and above 10% in Iowa and New Hampshire.

[hazek]

Ok but one national poll says 5% and one says 13%. Which one should we believe?

Forget the national, it's too early and practically irrelevant.

[John F Kennedy III]

No the polls aren't accurate.

[notsure]

In 2007 we also thought the polls couldn't possibly be the picture of reality, and yet they turned out to be remarkably accurate come election day.


I think that to some degree the polls paint this 'picture of reality'. Instead of reflecting voter intentions, they steer it. They give the average voters, who don't really pay attention to the issues, a compass for whom to direct their vote toward. I've never been polled. I've heard some people say, that when they were polled, that Ron Paul wasn't even mentioned as a candidate.

[Billay]

I see we've learned nothing over the last 4 years.

[hazek]

I see we've learned nothing over the last 4 years.

What can you expect from some fanatics around here?

[Galileo Galilei]

They were totally inaccurate with Rand Paul. Off by 11% on the average for the primary.

[sailingaway]

The question is when you get two national polls out on a given day such as PPP's saying Ron is at 5% nationally, and Reuters/Ipsos saying Ron is in third place nationally, at 13%. They both can't be right.


They were totally inaccurate with Rand Paul. Off by 11% on the average for the primary.

this^^

And PPP was one of the way off pollsters (don't remember the exact amount off) Rand's internal polls stayed pretty steady and PPP went way off on a Conway come back tangent, then pulled back to close to what actually happened the day before the election to preserve their rating. Only the polls within 2 weeks of the election apply to their rating. I put that together with the really weird selection of who they would poll head to head with Obama so long keeping Ron out when he was polling high and in states where he does well, and personally I think they have an agenda. I'm sure they stay within boundaries they can defend, but I think there is a lot of room within those boundaries.

[69360]

I looked at the RCP average just a minute ago. I think every one is accurate except the PPP poll. I am sure it has been manipulated.

[Verrater]

I remember Doug Wead saying way before he signed on to Ron's Campaign that polls aren't really reflective unless they reach about 1501 in sample size.
So I usually dismiss a lot of polls because of this.

[Peace&Freedom]

Those bringing up the 'accuracy' of the polls are missing the same point as in 2007-8. The polls are not objective players in the election game, they actively influence and massage the political coverage of campaigns and issues, and the momentum of candidates. I say this as someone who worked at Lou Harris. Most political polls are sponsored by the media, who frequently direct the polling firms to construct patently biased surveys that butter up certain issues or candidates, while burying others---then report the results as if it were neutral information.

The routine of the major media talking about candidate X 10 times more often than candidate Y strongly massages the poll numbers down of candidate Y, suppressing any natural upward momentum for that contender, which again makes the 'objective' results of the subsequent polls simply a mirror reflection of massaged coverage. The decision by the MSM to not include Paul in most of the polls in 2007 created and sustained his low poll numbers.

Those numbers, when carried over into the primary results, confirm only that the MSM knows how to manufacture defeat for candidates regardless of their individual merits or organization. Throw in the likelihood of voting fraud in key early contests to line up the results to fit the polls, and what we have is a almost totally manufactured election process in present day America.

[nf7mate]

Polls are accurate, but not foolproof. The more polls there are with the same result, the more reliable that result is (unless you believe it to be a big conspiracy, in which case there is nothing much else to say).

Polls can give valuable information, especially when they are broken down demographically or focus on a particular geographic area. Don't get mad at the thermometer when it's hot.

[InTradePro]

I remember Doug Wead saying way before he signed on to Ron's Campaign that polls aren't really reflective unless they reach about 1501 in sample size.
So I usually dismiss a lot of polls because of this.

^ this.
Basically you want a sample size of 2000 if you just asked a simple yes/no question. When you ask multiple choice questions (like which candidate you would vote for) you want even more in the sample. So all these polls are just a rough guide and should be taken that way. Also you can easy give guided result for example by listing Sarah Palin, Chris Christie as an option or adding not decided etc. All theose little things distort the results but sadly both people and journalist just read the headline spin just like any other reporting.

[jkr]

no

[robert9712000]

So do polls tell you what the people are thinking or telling the people what to think?

[hazek]

So do polls tell you what the people are thinking or telling the people what to think?

If later was the case then how come IF YOU CARED TO CHECK THE LINK TO 2007/08 DATA do we see polls showing McCain suddenly getting more and more support and eventually coming up on top. Polls don't tell people what to think, the media propaganda does, stupid!

Check the news reference graph for Dec 07:
http://www.google.com/trends/viz?q=John+McCain,+Mitt+Romney,+Ron+Paul,+Mike+Huc kabee&date=2007-12&geo=us&graph=weekly_img&sort=1&sa=N
http://www.google.com/trends?q=John+McCain%2C+Mitt+Romney%2C+Ron+Paul%2C +Mike+Huckabee&ctab=0&geo=us&geor=all&date=2007-12&sort=1

[TheTexan]

The statistics you listed show that they ARENT accurage.... the polls showed ~5-6% average, and the actual caucus results were 10%. In fact, I don't think there's a poll in there at all that showed > 10%, which further emphasizes that the polls are inaccurate.

And, again, in 2008, I loved Ron Paul but didn't vote at all because he had zero chance of winning. (Self fulfilling prophecy, I know, partly my fault) Only the "principled" voters voted for Ron Paul... the people who knew they'd lose, but voted to make a point.

This election, Ron Paul not only has "a chance" but he has a GOOD chance of winning. Nothing motivates people to go to the polls like a chance at victory.

[robert9712000]

You dont have to resort to calling me stupid when asking a question.IMO polls are a form of propaganda.Like saying Ron Paul cant be elected enough times ,that people decide not to vote for him because they feel he cant win thus being a self fulfilling prophecy and making him unelectable.

[TheTexan]

^ this.
Basically you want a sample size of 2000 if you just asked a simple yes/no question. When you ask multiple choice questions (like which candidate you would vote for) you want even more in the sample. So all these polls are just a rough guide and should be taken that way. Also you can easy give guided result for example by listing Sarah Palin, Chris Christie as an option or adding not decided etc. All theose little things distort the results but sadly both people and journalist just read the headline spin just like any other reporting.

Sample size is generally less significant than sample selection, at least in this election. The standard selection process is "Likely Republican Voters" meaning old people, and people who voted in the Republican primary in the past. Many Ron Paul supporters are neither.