The Candidates
Featured Archives
WASHINGTON — While public polls show Mike Huckabee leading Mitt Romney in Iowa, a new survey of an oversized sample shows Huckabee slipping and no longer ahead of Romney.A private corporate interest commissioned a phone bank survey of 15,000 Iowans who say they will attend Republican presidential caucuses Jan. 3. It showed Romney with 30 percent and Huckabee at 26 percent. Sen. John McCain was third with 12 percent and Rudy Giuliani fourth at 9 percent. Fred Thompson had only 1 percent, with slightly fewer votes than Rep. Ron Paul (also at 1 percent).
December 29th, 2007 at 1:34 am
Fred 1%? Would seem to call this poll into suspicion.
December 29th, 2007 at 1:34 am
Same regarding Ron Paul.
December 29th, 2007 at 1:43 am
This poll can’t be accurate – at least not for Fred’s numbers.
December 29th, 2007 at 1:46 am
Anyone know how this poll was conducted?
December 29th, 2007 at 1:50 am
1% for thompson seems a bit odd….
December 29th, 2007 at 1:52 am
And what was the supposed time frame of this “survey”…..
December 29th, 2007 at 1:54 am
I don’t like it….
I don’t buy it….
I want the Huckster to lead all polls into caucus night…..
December 29th, 2007 at 1:58 am
Iowans in general are way too smart not to come back home to Romney on Jan. 3. Huckabee’s recent idiotic Pakistan comments only underscore how lost he is on foreign policy.
December 29th, 2007 at 1:58 am
15,000 sample size seems odd… Though CW says that Romney and Huck are on par with eachother in terms of caucus support. It’ll all come down to whoever is more organized.
December 29th, 2007 at 2:00 am
Stats question–Sample size 15000 should increase confidence and reduce MoE % correct?
December 29th, 2007 at 2:01 am
Jason Peery, if Iowans are thinking about Pakistan and terrorism, they’re going to be gravitating away from both Huck AND Mitt.
December 29th, 2007 at 2:07 am
Metro,
Sadly, I don’t think there are very many people who know who Benazir Bhutto is, even if she’s been splashed on the news lately.
December 29th, 2007 at 2:09 am
OK, Fred will get more then 1%. Anyone want to bet me on that?
December 29th, 2007 at 2:09 am
Hah…The Huckster claimed John Bolton as someone who was helping him with Foreign Policy…..
Bolton says….I’m not, and I haven’t spoken with him…..
December 29th, 2007 at 2:18 am
Huckabee is a liar.
December 29th, 2007 at 2:20 am
Abe,
I thought that was so hilarious, I copied it and posted it. Does this guy even have a clue??
At a Thursday evening press conference, Huckabee said, “I’ve corresponded with John Bolton, who’s agreed to work with us on developing foreign policy.â€
Bolton, however, has a different view. “I’d be happy to speak with Huckabee, but I haven’t spoken with him yet,†said Bolton, now a senior fellow at the American Enterprise Institute, a conservative think tank in Washington.
December 29th, 2007 at 2:21 am
#10
In theory it should. But 15,000 is a freakin’ huge sample size for such a small population. I can’t imagine they called up that many real caucus-goers.
December 29th, 2007 at 2:21 am
Where is the MSM on that….or does Mitt only get attacked for his comments, even though 4 Historical sources state george Romney actually walked with MLK
December 29th, 2007 at 2:23 am
In regards to Iowa Caucus…isn’t 15,000 the average number that attends the caucus? Could that be the reason for such a large number??
December 29th, 2007 at 2:28 am
Jason, thanks for the info. Can you give somemore info on this poll?
December 29th, 2007 at 3:02 am
Thomas (#17) Especially with all of those calls from all campaigns, polls and 527s, I will be surprised if there are that many people still willing to take one more phone call, especially around the holidays, when they are in midst of their out-of-town families!
December 29th, 2007 at 3:34 am
Isn’t 15,000 something 20% of the total number of people that will caucus? That is a huge sample size. They should have just kept going and called them all.
December 29th, 2007 at 6:58 am
No, they are expecting something like 90,000 will caucus on the Republican side
December 29th, 2007 at 8:28 am
This poll is B.S. Fred is not at 1 %.
The reality=, confirmed by all recent polling, is that Huckabee is ahead by a few points. Unless or until that changes, Romney is in deep trouble.
December 29th, 2007 at 8:46 am
Fred will place no lower than third in the caucus. So this poll is crap, just like the rest of what Robert Novak posts. Ron Paul at 1%? Fred at 1%? Crap.
December 29th, 2007 at 9:15 am
Fred will place no lower than third??? Oh pleeaaasseee! And Ron Paul has his teckies travel and follow him around, but Iowa Caucuses don’t work that way. Good News, only 5 more days and all the speculation will be laid to rest.
December 29th, 2007 at 9:36 am
No way is Fred over 0%. This poll is flawed.
December 29th, 2007 at 9:37 am
Does someone want to tell me why the hell Romney is now below keyes and hunter on the rankings? Or for that matter, why Giuliani still leads?
As for this poll, I don’t know, but we will see what happens. DMR comes out tomorrow, correct?
December 29th, 2007 at 9:41 am
The question is how ‘representative’ is the sample of 15000 voters? The more the merrier is not valid in statistics.
December 29th, 2007 at 9:43 am
#28. Rankings became alphabetical order? — when in doubt, go alpha.
December 29th, 2007 at 9:53 am
Mitt is going to win the caucuses because of organization and the vetting of Huckabee. I believe that this will be verified by the Des Moines Register poll tomorrow. So much has been coming out about the Huckster that there is no way to keep his support intact. The numbers for the top 3 guys in this survey all seem about right to me. I seriously doubt if Giuliani will poll this high and, of course, don’t believe that Fred will poll this low. I will say this, however….Fred voters in a 2-man race in Iowa will actually tend to vote for Mitt. It’s the rational thing to do when the alternative is Huckabee….and Fred is running a campaign geared to rational people.
December 29th, 2007 at 10:05 am
I’d like to know if any of the Huckabee people on here have changed their mind in the last couple weeks. It would seem to me that the people on here would be changing their mind even more than the average Joe on the street. If any of you have a conversion story, please share it.
December 29th, 2007 at 10:05 am
Rookie here but Novak doesn’t normally put out junk–at least limbaugh thinks as such.
December 29th, 2007 at 10:17 am
I have a conversion story. I used to dislike redneck hick republican supports that don’t understand fiscal policies, now I have converted and dislike them even more after hearing their reasons for support huckabee.
December 29th, 2007 at 10:18 am
* supporters *
December 29th, 2007 at 10:20 am
Carl,
“Rookie here but Novak doesn’t normally put out junk–at least limbaugh thinks as such.”
True. But Novak’s record of success is decidedly mixed. Just because he says “Huckabee is slipping” doesn’t make it so – and certainly basing such a claim on a single suspicious poll is a little too ballsy. Novak is the same guy that told us we’d have 54 GOP senators this cycle.
December 29th, 2007 at 10:25 am
#28 They’re in alphabetical order with no rankings. How convenient for Kavon
December 29th, 2007 at 10:28 am
#30 Colin
You’re kidding right? Stats is all about sample size and how accurate of a predictor based on how small of a sample… So if they polled 100% of Caucus goers, that wouldn’t make it more relevant according to your logic.
December 29th, 2007 at 10:29 am
Oops #29 for comment #38.
December 29th, 2007 at 10:31 am
Carl Cameron (Fox News)said we should not trust in polls these days, its weekend, next monday is new year’s eve, tuesday is new year’s day. The only thing clear is the race is close.
December 29th, 2007 at 10:33 am
Carl Cameron is a goofball – I don’t believe anything he says and wish he’d go back to some local tv station. Go watch that debate and see what a self-promoter he really is. Give us a break Carl.
December 29th, 2007 at 10:34 am
#29 – WHAT???? Do you know anything whatsoever about what larger sample sizes to to the margin of error???? Apparently not! Given that people were telling the truth about who they would vote for and if they are really going to the caucus, the MOE here would be extremely low, less than 1/2 of 1%. Whether or not they were telling the truth is unknown, but the same can be said in the small little sample sizes.
December 29th, 2007 at 10:36 am
This new ranking is crap, just admit the true: Giulini is in 4th, McRino is 3, Nutsckabee is 2 and Romney (until thursday) is #1.
December 29th, 2007 at 10:36 am
Carl Cameron is very down on Mitt Romney, and always has been.
December 29th, 2007 at 10:38 am
42 – 29 is right – statistically speaking
December 29th, 2007 at 10:39 am
Carl was talking about the LA Times/Bloomberg poll that has Nutsckabee with more than 10 points over Romney.
December 29th, 2007 at 10:47 am
What???? No way!!! He’s not right at all!! I spent a ton of years in statistical process control, and the larger the sample size, the more accurate! Period. Anyone who thinks otherwise, better go back to their statistical basics.
December 29th, 2007 at 10:51 am
The only way it wouldn’t be more accurate is if the length of time over which the data was gathered caused the answers to not capture the current situation as well. If this took 10 days, for example, it could be less accurate because of change of minds. Other than that, it is very, very, accurate.
December 29th, 2007 at 10:51 am
Does anyone have insight into the tallying of the votes on caucus night – I mean, how exactly does it take place?
Here’s what I think – please correct if you have more insight:
- GOP caucus goers arrive at 7pm at their caucus location
- Organizational remarks, maybe brief spills trying to woo some undecideds
- Maybe around 7:45 or so the attendees will write down their vote on a blank piece of paper and turn it in
- Caucus location captains/team (?) will tally the results of the poll which could take a few minutes to over an hour perhaps (?)
- Each precinct submits their tally to the state GOP (?) – phone?/fax?/email?/courrier?
- All of the precinct numbers are aggregated and published to the media
Questions:
- What time constraint are the caucus goers under to vote? ie. someone arriving at 8:30 – can they still participate?
- What time guidelines are in place for each precinct to report totals?
- What expectations do the media have on actual totals? (ie. w/70% of locations reported, here are the results…; with 98% of locations reported, etc…)
So when you add this all up, are we looking at an 11pm ET news cycle? Or will the results not be clear until the next day?
Any insight/clarification would be great. Thanks
December 29th, 2007 at 10:54 am
#49 – You forgot opening prayer!
December 29th, 2007 at 11:00 am
Manufactured news from the Romney camp. Novak is a whore.
December 29th, 2007 at 11:05 am
Huckabee 08 All Rhetoric no substance all the time
December 29th, 2007 at 11:12 am
Keep an eye on the intrade Iowa betting odds today. Currently Huck 60.1 Mitt 35
December 29th, 2007 at 11:27 am
Paul S – Novak’s a guy, right?
December 29th, 2007 at 11:29 am
#38. Shawnie, I am not kidding. You need to study the terms “bias”,”random” and “representative sample”.
So if they polled 100% of Caucus goers, that wouldn’t make it more relevant according to your logic.
If they polled 100% of Caucus goers, it will not be called a sample.
December 29th, 2007 at 11:48 am
Shawnie, you’re right, my number 47 & 48 posts are correct. And, I didn’t mention that if it were taken over a longer period of time, that just means that the other polls were overstating Huckabee’s lead all along, and with Mitt’s upward movement relative to Huck lately, he could even be up by more than that shown here. Any statistician on here that doesn’t agree is in a state of denial.
December 29th, 2007 at 11:55 am
Illinoisguy – I understood Colin to be referring to the polling parameters and sampling statistics – not simply the sheer number of responders. If the underlying parameters are not reasonable, a larger sample size does not equal a more accurate poll. To use a religious stereotype – if the pollsters sample was broken down 90% mormon, 10% evangelical, the results would be flawed, regardless if they polled 500, or 10,000 as the 90/10 are not accurate population/voter figures.
December 29th, 2007 at 12:01 pm
#57. Right you are.
December 29th, 2007 at 12:12 pm
This poll, until proven otherwise, is crap. Not just because of the numbers. They polled 15000 likely voters? How in the hell did they get that type of turnout for a polling data unless they polled over a month? Are they saying that they were able to contact 15000 likely republican caucus voters over December 24-26? Was it commissioned over a couple of weeks? What were the questions asked? Was it a robo phone poll?
I do think the race is tightening between Huckabee and the rest of the field, but Novak seemed to imply in the interview with Hannity yesterday that this was recently taken, but in the days after the news and newspapers have been swamped with coverage of the Bhutto assasination, what factor would that play? How would they contact 15000 people in that short amount of time. I’m no statician and am aware that larger samples mean less MoE, but I did take statistics and do know that it doesn’t mean its accurate. Also, there are scientific factors that play when using such a large amount of data. There’s just no inconcievable way that a polling firm could perform that type of poll within a matter of relevence.
This is like basing results on an internet poll, which is what it suspiciously sounds like. I’ve always been a big Novak fan, but he’s also the first to admit that when it comes to predicting electoral outcomes, he has a bad track record.
December 29th, 2007 at 12:28 pm
2 things:
A larger sample size would tend to be more representative of the population as a whole. There is no reason to think that adding to sample size would create MORE of a sampling problem. In fact, to think this would be counterintuitive and is most likely a red herring.
Additionally if the survey were done over the last two weeks, then this is EXTREMELY good news for Romney as data from two weeks ago was definitely not in his favor, meaning that the current trend is very positive.
If the private investors in this poll were Romney backers, it would be understandable why they wouldn’t release the poll publicly. Let team Huckabee feel that they have a seven point lead, don’t give them extra motivation to turn out. Huckabee’s voters will mainly be emotionally driven. Don’t back them into a corner.
December 29th, 2007 at 12:28 pm
59 – I suspect it was a robo phone poll with touched response. Probably some qualifying responses before getting into the ‘press 1 for Huckabee, press 2 for McCain, etc..’ 15,000 robo calls could be done in a matter of hours depending on how many lines they have dedicated to the effort. A lot of the statistical noise from robo calls would balance out with that large of a sample (ie. kids answering the phone and screwing with the poll, etc.)
I think Novak’s comments speak more to the tightening of the race in IA and less about what the actual spread is and who’s in 1st or 2nd. I wouldn’t set this report aside completely, but Romney supporters wouldn’t want to break out the caffeine free root beer just yet either.
December 29th, 2007 at 1:07 pm
i agree with 60 and 61 both have good points