How accurate were 2008 polls at predicting Ron Paul’s support?

[jeremiah89]

Here on the forums and elsewhere wherever Ron Paul supporters hang out, there seems to be a healthy skepticism about how accurate the polls are, and if they're maybe way underestimating Ron Paul's support.

However, I just saw this article that compared 2008 polls to actual results. If anything, other than Iowa, it appears his support was overestimated.

http://www.capitalfreepress.com/3866-accurate-2008-polls-ron-paul-support
(http://www.capitalfreepress.com/3866-accurate-2008-polls-ron-paul-support/)

All of these states except for Iowa held primaries rather than caucuses, which perhaps explains Iowa being the only state where polls significantly underestimated Ron Paul’s support. In the primary states, the polls pegged his support within 1% and in everywhere but Michigan slightly overestimated his support rather than underestimating it.

Just thought this would be good to pass on. Everyone should keep up the great work, but don't become too confident. :)

[kylejack]

Yes, they were roughly accurate. If we win early states, we have the ability to shift the vote in future states.

[69360]

They are a little low, maybe a few points. Anyone who thinks we are winning right now nationally is misguided.

[sailingaway]

It's how inaccurate they were at predicting Obama's. All the action and independents were on the Dem side last time. Our problem is the question of how much will the winter break impact the student vote, since they tend to take vacations then, and won't be peer supporting each other.

Right now there are two sets of polls, the ones that only poll those who consider themselves Repulicans, where Paul does less well, and the ones polling all who intend to caucus in the GOP caucus, regardless of their current party (since you can change party right at the caucus.)

edit -- the above was regarding Iowa only. Ron isn't really known in the other states and needs good enough showing to get the attention of voters in later states.

[RonPaulFanInGA]

Having been here in 2007, the polls were a big conspiracy and an army of cell phone-using 18-29 year olds were going to sweep Ron Paul to victory in the GOP primary.

It's sad to see so many not learning from history, and openly dismissing every poll simply because the result is not liked.

[ronpaulitician]

The polls were pretty accurate at predicting Ron Paul support in early states.

[hazek]

The last two weeks before the election they were very accurate:

See for yourself: http://en.wikipedia.org/wiki/Statewide_opinion_polling_for_the_Republican_Party _presidential_primaries,_2008

[Agorism]

Last DMR poll in Iowa at him at 9% or 10%, I forget, and he got 11%.

[tsetsefly]

THey where within the margin of error.

Polls are usually accurate, now is it poll's that lead people to vote a certain, that is another argument but we can't rely on bad weather, poll bias etc. The poll numbers have to improve for us to have a chance. I think if we get within the margin of error of 1st place we can win Iowa. After that it gives Paul's candidacy a huge credibility...

[justatrey]

Having been here in 2007, the polls were a big conspiracy and an army of cell phone-using 18-29 year olds were going to sweep Ron Paul to victory in the GOP primary.

It's sad to see so many not learning from history, and openly dismissing every poll simply because the result is not liked.

This. Although I believe Rasmussen consistently has Paul doing considerably worse than the other polls. So imo Rasmussen polls understate his support, while the average of the other polls is usually pretty close but maybe 1 or 2% lower than what he ends up getting.

[Jingles]

Honestly, I think the polls are going to be a little less accurate this time around. I can barely think of anyone I know (and these people are between the ages of 18 - 70) that own landlines anymore. My grandparents on my mom's side and my dad's side completely got rid of their landlines. They just use a cellphone now as their primary phone. Same with a lot of my friends' parents and etc... I'm sure they will be semi-close, but I don't know. I can only speculate. We won't really know until after Iowa and NH.

[Galileo Galilei]

Here on the forums and elsewhere wherever Ron Paul supporters hang out, there seems to be a healthy skepticism about how accurate the polls are, and if they're maybe way underestimating Ron Paul's support.

However, I just saw this article that compared 2008 polls to actual results. If anything, other than Iowa, it appears his support was overestimated.

http://www.capitalfreepress.com/3866-accurate-2008-polls-ron-paul-support
(http://www.capitalfreepress.com/3866-accurate-2008-polls-ron-paul-support/)


Just thought this would be good to pass on. Everyone should keep up the great work, but don't become too confident. :)

They were very inaccurate [rigged]. In Iowa, 56 polls underestimated his support, and two got it right. Not a single poll overestimated his support! At the time of the Iowa straw poll, Paul was polling at 0%, yet he got 9% in the straw poll.

In NH, there were 84 polls. 10 overrated Paul. 18 got it right. 56 underrated Paul's support!

As for Rand Paul, even the final polls were WAY OFF. In the primary, the final polls were off by 11%!

S*** like this does not happen by accident!!

source here: http://www.realclearpolitics.com/epolls/2008/president/gop_results.html

[Galileo Galilei]

So to summarize the 2007/2008 polls in Iowa and NH regarding Dr. Paul:

110 underrated his support.

20 got it right.

10 overrated his support.

[da32130]

Here on the forums and elsewhere wherever Ron Paul supporters hang out, there seems to be a healthy skepticism about how accurate the polls are, and if they're maybe way underestimating Ron Paul's support.

However, I just saw this article that compared 2008 polls to actual results. If anything, other than Iowa, it appears his support was overestimated.

http://www.capitalfreepress.com/3866-accurate-2008-polls-ron-paul-support
(http://www.capitalfreepress.com/3866-accurate-2008-polls-ron-paul-support/)


Just thought this would be good to pass on. Everyone should keep up the great work, but don't become too confident. :)

1) I've covered this here before:
http://www.ronpaulforums.com/showthread.php?328197-How-Ron-Paul-Wins

caucus turnout can have a big impact:

"Last time Iowa was also much more competitive. Romney was putting in a huge effort. A lot of voters (that are dispersed now) went to Huckabee. In early caucus states where the effort wasn't as great by other candidates Ron Paul did much better than polling showed(shortly after these results we only got 4% in the Virginia primary, difference was turnout):

22% in Washington
21% in North Dakota
25% in Montana
16% in Minnesota
17% in Alaska
18% in Maine
14% in Nevada (2nd place finish)
10% in Iowa (highest poll was 10%, but average at time of caucus was only around 7% - he is polling consistently in the 12% area now, which could mean high teens support, if not more - that is why phone from home, etc. matter, increasing turnout ensures victory)"


2) How the youth vote pushed Obama to victory:
http://www.ronpaulforums.com/showthread.php?332053-IOWA-2008-Exit-Polling-and-2012-current-polling

Only one pollster didn't understate Obama's margin of victory. And only one pollster didn't understate Paul's Iowa result last time.

[Keith and stuff]

So to summarize the 2007/2008 polls in Iowa and NH regarding Dr. Paul:

110 underrated his support.

20 got it right.

10 overrated his support.

That is an interesting way to look at it. In NH, the final 2 Rasmussen polls overestimated Paul's vote but most of the other final polls were pretty accurate. Some were a little high and others were a little low, for the most part. So as for NH, the polls were great, overall.

The final polls were inaccurate in NV. However, it is harder to poll for a caucus because people are more likely to change their mind while caucusing.

[danbeaulieu]

There's a dynamic that I think is important... The young Ron Paul votes in 2007 were greatly diminished by Obama's fake message. This time around there is no Obama to take votes away from Paul in the primaries. Those young people aren't recorded accurately, so hopefully they account for a couple of percentage points bump.

[brandon]

The polling companies are free market ventures that do their best to achieve accurate results. There's no reason not to trust them. Sure some of them are wrong, but they aren't any more wrong about Paul than they are about the other candidates.


Paul is in a great place and I think he could see a surge from the most recent debate that puts him very close to 1st place in Iowa.

[bbartlog]

There were also problems with the GOTV effort in 2007. Between that and the greater likelihood of independent support in Iowa, I think you can assume that the polls in Iowa will be off on the low side. And in general, the turnout for caucuses is low enough that polling those states is hard, the margin of error large, and the opportunity for enthusiasm and organization to matter very significant.
For states with primaries, though, the polls are usually pretty close.

[PMix32]

Honestly, I think the polls are going to be a little less accurate this time around. I can barely think of anyone I know (and these people are between the ages of 18 - 70) that own landlines anymore. My grandparents on my mom's side and my dad's side completely got rid of their landlines. They just use a cellphone now as their primary phone. Same with a lot of my friends' parents and etc... I'm sure they will be semi-close, but I don't know. I can only speculate. We won't really know until after Iowa and NH.

There are several pollsters who use random number generators from which to draw their polling lists. Thus hey are equally likely to poll any number whether they are cell phones or landlines, registered or unregistered. They downside to this is that people who have both cell phones and landlines are slightly more likely to have one of their numbers come up. People who share landlines will be less likely to be polled because the number is equally likely to be chosen, but only one person can answer it. Moving closer and closer to every adult using exclusively a cell phone will only help the good pollsters to get more and more accurate. If everyone had their own personal cell phone, then there would be a perfect correlation of one phone number and one person.

[Galileo Galilei]

The polling companies are free market ventures that do their best to achieve accurate results. There's no reason not to trust them. Sure some of them are wrong, but they aren't any more wrong about Paul than they are about the other candidates.


Paul is in a great place and I think he could see a surge from the most recent debate that puts him very close to 1st place in Iowa.

That's not what happened with Rand Paul in the primary. Every single poll underrated his actual support except one poll released on the eve of the election when it was too late to have influence. The polls were off by an average of 11% per poll!