Landmark Communications/Rosetta Stone Georgia 2012 GOP Primary Poll
- Newt Gingrich 38% (35%)
- Rick Santorum 25% (26%)
- Mitt Romney 19% (16%)
- Ron Paul 5% (5%)
- Undecided 13% (18%)
Survey of 1,300 likely Republican primary voters was conducted February 23, 2012. The margin of error is +/- 2.7 percentage points. Results from the poll conducted February 9, 2012 are in parentheses
–Data compilation and analysis courtesy of The Argo Journal
February 25th, 2012 at 12:27 am
Grinch wins “home state,” Santorum doesn’t take so well in south…splitting momentum and letting Mitt take the majority of Super Tuesday states. Loving the way it’s turning out at the end of this week. Can’t wait for…
TEAM MITT’S TWO VICTORY TUESDAY ON THE 28TH!!!
February 25th, 2012 at 12:30 am
Speechless on Stone Mountain.
Give em severe hell “Peachtree” Mitt!
February 25th, 2012 at 1:06 am
Things are looking good for Mitt on Tuesday.
Things are looking not so good for Mitt on Super Tuesday.
At this time, there are three Super Tuesday states for which corroborated multiple polling exists, Georgia, Ohio and Oklahoma. And Mitt’s not winning any of them.
Here’s the current Georgia CPSR picture, based on the two latest polls taken from 2/22-23 by Rasmussen and this latest release from Landmark here, with trend indicators based on the two previous surveys taken by Survey USA and Mason-Dixon between 2/1-8:
GEORGIA Composite 2/22-23 Final
1. Gingrich 35.3-37 -3.3-11.7
2. Santorum 24-27.7 +10.3-19.7
3. Romney 16.3-21.7 -5.6-16.7
4. Undecided 10.3-14
5. Paul 5-7.7
As you can see, at the moment, it looks like a narrow Gingrich victory with Santorum close on his heels, the two together claiming the majority of votes. Romney looks headed for a weak also-ran third place finish here.
The next Super Tuesday state for which we have corroborated multiple polling is Ohio. Here’s the current Ohio CPSR picture, based on the two latest surveys taken 2/7-15 by Quinnipiac and Rasmussen, with trend indicators based on the previous survey taken by Quinnipiac on 1/9-16:
OHIO Composite 2/7-15 Final
1. Santorum 38-40.2 +14.8-25.4
2. Romney 24.8-28
3. Gingrich 15.8-17
4. Paul 6-13.2
5. Undecided 7-10.2 -8.6-20.2
As you can see, currently Romney isn’t winning here either; he is scoring a close second behind Santorum, with Santorum probably in the upper thirties and Romney probably somewhere around 25%. The two together are receiving support from over half of the primary electorate. No one else is making it to that top tier.
The third Super Tuesday state for which we have corroborate multiple polling is Oklahoma. Here’s the current Oklahoma CPSR picture, based on the last Sooner and Rasmussen surveys taken between 2/8-21, with trend indicators derived from the previous Sooner poll, taken 11/17-12/16:
OKLAHOMA Composite 2/8-21 Final
1. Santorum 39-44.7 +32-44.7
2. Romney 17.3-22
3. Gingrich 18-23.7 -4.4-19.9
4. Paul 3-11
5. Undecided 7.3-13 -6.1-21.6
Romney and Gingrich are statistically tied, but since Gingrich has trended down from the previous survey whereas Romney is steady, I give Romney the edge and second place to Gingrich’s third place. Similarly Paul and Undecided are statistically tied but, since Undecided has trended down since that previous Sooner poll, I place Undecided beneath Paul.
And once again, as you can see, Romney is not currently headed for a win here. Rather surprisingly, given this is a Southern state where most believed Gingrich would do well, it is once again Santorum rather than Gingrich who seems to be exciting conservatives, even here in so-called Gingrich territory. In fact, Gingrich does not even appear to be running second here. This is illustration, if illustration were needed, of how weak Gingrich appears to have become even with his own base. He’s coming in a weak third behind Romney, with Santorum coming on very strong in first place in the low forties compared to his nearest rival, Romney, who only manages to score a distant second around 20%.
By the way, Killjoy suggests these updates of mine would be ideal on Twitter. I’ve got a Twitter account, but I’m not sure what hash tag to use if I start posting abbreviated versions of this sort of stuff on Twitter. Any neat ideas for a hash tag? Thanks.
February 25th, 2012 at 1:48 am
i think newts a slam dunk in georgia, the race for 2nd will be big.
oklahoma will go to santorum im confident.
Ohio will be all about how much momentum mitt gets coming out of Michigan win.
if he gets a bunch of MO then he has a chance. otherwise im thinking he barely loses.
no data, just gut feelings…
February 25th, 2012 at 1:55 am
Is is true that if Mitt runs third in many ST (southern) states he won’t get any delegates? That sux
.
February 25th, 2012 at 2:27 am
5:
This poll shows Romney at 23% and Gingrich at 18% in Oklahoma
http://soonerpoll.com/santorum-leads-in-oklahoma/
Also using 2008 primary ideology distribution to Rasmussen cross tabs made it instead of Gingrich leading 22-18, a tie at 20-20.
In insider advantage Georgia poll had Gingrich 26, Romney 24, Santorum 23. In Rasmussen cross tabs “certain of who they vote” Santorum is at 25, Romney 21.
In 2008 Romney, McCain and Huckabee were all within 4 points from each other’s and Romney won some Atlanta CDs.
Even when Santorum led in Michigan, he overwhelmingly lost in “Best chance to beat Obama” cross tabs in Rasmussen poll, when he was exposed, his numbers collapsed, his point was his purity, him being “true conservative”, if you have to choose between two non-pure candidates, might as well go for the more electable. Santorum is in the “Best chance to beat Obama” cross tabs 12 points behind Romney and 11 points behind Gingrich in Georgia, expect the debate, and losing Michigan, Arizona and Washington to collapse his numbers.
I’m pretty sure Romney will get delegates from many southern congressional districts, being second, and the winner being below 50% in CD (Georgia, Oklahoma), or in Tennessee’s case below 67%, is enough for 1 CD delegate.
February 25th, 2012 at 6:30 am
Thanks T!
February 25th, 2012 at 6:42 am
Is there any chance Santorum (or Gingrich) drops out if Romney does very well
in AZ and MI?
Super Tuesday looks like a mixed-bag where everyone (except Grand Wizard Ron) gets
to claim a victory. So that won’t resolve anything unless Gingrich loses GA.
Without Rick or Newt dropping out, this contest will easily drag on into April, depleting
resources (Obama has $76 million cash on hand, ten times as much as Romney), creating more
opportunities for gaffes, etc. And frankly, pulling the candidates even more to the right.
I am convinced Romney would not have changed his econ plan if he was not in a fight for
Michigan.
February 25th, 2012 at 7:30 am
If Mitt can get some momentum and take his own voters back from Santorum by Super Tues, Mitt will come in 2nd in GA.
February 25th, 2012 at 8:24 am
9- Zactly so!
February 25th, 2012 at 8:30 am
Gingrich has a foothold.
February 25th, 2012 at 8:50 am
Mama:
It’s called foot in mouth.
Give em severe hell “Hotlanta” Mitt!
February 25th, 2012 at 8:50 am
Mass Con – I was just going to express the same sentiment abut 2nd place.
Which means the delegate count keeps growing for Mitt.
The AZ primary is being given short shrift by the media due to the horse race/home state angle of Michigan
BUT
Az has 29 delegates and is WINNER TAKE ALL I believe
February 25th, 2012 at 9:12 am
13
Yep.
February 25th, 2012 at 10:01 am
Delegate math favors Romney though.
Virginia – 3 delegates per CD won, if one candidate gets over 50% of the statewide vote, all at-large delegates go to him, and since only Romney and Paul are on the ticket, this will most likely be de-facto winner-take-all for Romney – 49 delegates (46 delegates, 3 unbound RNC delegates, would assume going to Romney)
Idaho – 25% LDS, caucus math enables winner-take-all for Romney easily (32 delegates)
Vermont – if Romney gets over 50% as it is likely WTA for him – 17 delegates (14 delegates, 3 unbound RNC, most likely will go to Romney)
Massachusetts – delegates proportionally to everybody getting over 15% of statewide vote, starts to look like that might mean WTA for Romney or over 80% of delegates:
This Suffolk university poll is from 2/11 – 2/15, so from time of Santy’s peak popularity and and his social conservative positions and the way he represents them might costs him some votes in MA. So highly likely that either Romney takes all delegates, or Santy barely manages to get over the 15% and Romney takes over 80% the delegates. So 41 delegates to Romney (38 bound, 3 RNC, can’t see them going to anybody else but Romney)
Romney 64%
Santorum 16%
Paul 7%
Gingrich 6%
Undecided 6%
Refused to answer 1%
Alaska District Conventions: Delegates allocated purely proportionally, significant LDS numbers give Romney slight edge, but because it is pure proportional, I would expect no matter who wins it Romney, Santorum and Paul each getting somewhere between 5 to 9 delegates, so in delegate math, no one is going to win significantly. (24 bound delegates, 3 RNC delegates). Ron Paul might be able to win this caucus I think, but no matter who wins it, Romney, Santorum and Paul are going to get somewhere between 6 and 10 delegates each. 28 delegates total.
North Dakota:
Romney won last time but really but not because of of LDS vote, not many LDS there. But it’s not like the “front runner” candidate does bad here either, on 2008 Super Tuesday, Romney got 8 here, and McCain 5 delegates. The delegates will be allocated proportionally, no separate precinct delegate selection like in IA, CO, MN and ME, so no delegate ninjaing here.
The South:
If the winner doesn’t win congressional district by 50% in Georgia or Oklahoma or by 67% in Tennessee, the CD delegates are split 2 to winner, 1 to second. Also half of the three southern state’s total delegates are divided based on statewide popular vote winners, the minimum required percentages are so low that Santorum, Gingrich and Romney will in all likelyhood pass them in each of the three southern states.
Winning MI, AZ and WA should give Romney enough momentum, and hurt Santorum, enabling Romney to come second in most CDs in Oklahoma and Georgia, maybe even winning some of urban/suburban CDs he won in 2008 in GA, Santorum wins most Oklahoma CDs, Gingrich most Georgia CDs. In Tennessee Romney wins at least 2 CDs, maybe 3, Gingrich might win some, Santorum some.
Actually I did some calculations that if candidate A was to perform solid first all around the South, winning almost all CDs becoming second in few, but without getting those required majority/super majority victories for WTA to hit, and candidate B would be second all around the South, winning some CDs, and then if we had candidate C who doesn’t win CD’s nor become really second either but gets enough popular vote to get on the proportional delegates, the candidate A would receive slightly less than half of southern delegates and candidate B would receive slightly over third of the southern delegates.
Gingrich doing well in Georgia, Santorum doing well in Oklahoma, Romney getting possibly boost from MI, AZ and WA (last Wyoming county caucus straw poll result will be released on Feb 29, but winning that straw poll which is going to based on results this far have less than 2k participants is probably not going to give that much momentum), becoming second most the south, winning some Tennessee CDs possibly some GA, that would give Romney, Santorum and Gingrich each about third of the Southern delegates.
Ohio is really nice price, 3 delegates for each 16 CD’s won, 48 CD delegates total. 15 at-large delegates proportional to those getting over 20% of statewide vote, seems like Romney and Santorum might be only candidates breaking that limit. Santorum isn’t even on the ticket in 3 CDs, so that will be most likely 9 delegates to Romney at least. If Romney manages to take Ohio with momentum from MI, AZ and WA and by exposing Santorum, this can be really nice delegate price for him.
So summary of super tuesday:
4 WTAs for Romney ID,MA,VA,VT – 139 delegates
2 proportional caucuses (AK,ND) – 55 delegates, Romney, Santorum and Paul probably get about equal amount of these delegates, Gingrich might get few also.
3 southern states (GA,TN,OK) – 177 delegates, most likely will be split pretty evenly to Romney, Santorum and Gingrich.
Ohio – 66 delegates, Santorum isn’t even on the ticket in 3 CDs, if Romney manages to take the momentum here from MI, AZ and WA, reveal Santorum as non-pure “true conservative” so that people go for the more electable, this can be quite a delegate boon for Romney.
February 25th, 2012 at 10:11 am
Super Tuesday polls are no as relevant until the week before voting day.
February 25th, 2012 at 10:16 am
15 to contineu:
So with those 4 WTAs, and 2 caucuses and the South splitting the way I suggested, an Ohio sweep with 60 delegates from there would give Romney almost 65% of Super Tuesday delegates, Ohio spanking with winning only 3 of the CDs where Santy isn’t on ticket and some of the proportional delegates would still give over 50% of Super Tuesday delegates to Romney.
February 25th, 2012 at 10:28 am
Teemu,
Thanks for the summary! Just in the 3 Southern states and Ohio, it doesn’t look like Mitt will do that well, but he will get delegates in each state, and combined with the delegates he gets from the other 6 states, he should get most of the delegates that day by a comfortable margin.
February 25th, 2012 at 10:38 am
American Research Group poll of Arizona:
Romney 39 (38)
Santorum 35 (31)
Gingrich 11 (15)
Paul 9 (11)
Brackets Feb 14-15
Poll conducted Feb 23-24 right after the debate.
These results are completely different from Rasmussen who yesterday showed it as a 13 point blowout for Romney.
Note while Newt dropped 4 points, Santorum gained 4 points.
Could the difference in the poll results be attributed to Rasmussen polling literally right after the debate while ARG polled a day of two after the debate once the dust had settled and voters had a chance to digest the Rasmussen results and/or had time to think about what just happened rather than respond in a knee-jerk fashion?
February 25th, 2012 at 10:38 am
Interesting ARG Arizona survey just out at http://americanresearchgroup.com/pres2012/primary/rep/az/ . While it still shows Romney with a narrow lead, it also shows that Santorum has moved up as compared to the previous poll-set.
In the current poll-set of two surveys, one from Rasmussen taken 2/23 and this one from ARG taken 2/23-24, Romney’s CPSR is 38-43% and Santorum’s is 31-33%. This compares to the previous poll-set of four surveys, from Rasmussen (2/16), PPP (2/17-19), CNN (2/17-20) and NBC/Marist (2/19-20), where Romney’s CPSR was 39.5-40.5% and Santorum’s was 28.2-30.5%.
In other words, while Romney’s current CPSR overlaps his previous CPSR, indicating that he is stable but not moving up or down, Santorum’s, on the other hand, has moved up a statistically significant amount between the 20th and the 24th, between .5 and 4.8 points. He is still behind Romney; if the primary election were held today he would probably lose. But Santorum is the one with momentum in Arizona, not Romney.
Of course, as some of you know, I don’t think we have true post-debate polling (the one where Santorum did so badly) until the polling from yesterday. I’d be very interested to know whether ARG saw a downturn in Santorum’s numbers on the 24th, but they haven’t revealed their daily figures.
Hope there’s a FPP post here soon about this survey.
February 25th, 2012 at 10:45 am
#20:
The bottom line in both Michigan and Arizona: How many more Gingrich supporters are willing to move over to Santorum to stop Mitt Romney?
Mitt got his initial bounce right after debate, but did the Rasmussen polls in both Michigan and Arizona scare the s*it out of the anti-Romney crowd that it has decided to pull out all stops to prevent Romney from winning?
February 25th, 2012 at 10:46 am
20
Criggs, you do an excellent job of over-analyzing crappy polls. ARG is on auto-trash as far as I’m concerned.
Remember Florida? I nailed my prediction, while yours, which put equal faith in good pollsters and bad, flunked.
Trust me, the movement toward Romney is real, and the movement away from Santorum is even more real.
February 25th, 2012 at 10:50 am
In the latest Rasmussen Arizona poll, Romney wins head-to-head against Santorum 52-40.
In Michigan poll also, despite it’s severely “very conservative” sample, Romney loses to Santorum only by 2 points, 44-46, if the very conservatives were not so much oversampled, Romney would win that also.
February 25th, 2012 at 10:57 am
22. As you know, if I see a poll which agrees with other polls, I presume it is therefore good. I believe that’s a difference between my methodology and yours. In this case, ARG’s results today agree with Rasmussen’s from 2/23, and therefore I judge them to be legit. I also think it’s hazardous to presume, other things being equal, that a poll is more likely wrong than right. I presume the opposite, as long as a poll agrees with other polls.
In the case of Florida, you got it right and I got it wrong. I presumed that Insider Advantage and PPP were both right, since they agreed with each other; but it turned out they were both wrong. In the long-term however, I’ve found my system is more likely to be correct than incorrect, as long as I have polling from the day before the election; for example, I WON the South Carolina contest, as I recall.
The bottom line is GIGO. We all can get burned by bad data; nature of the beast.
As for the movement toward Romney, even if one removes this ARG survey, I still don’t see evidence of a surge toward Romney in the Copper State. Since the February 1st Rasmussen survey, Romney has trended down between 3.5 and 12.5 points while Santorum has trended up between 11.2 and 21.5 points.
Whether one includes or excludes this latest ARG survey, the preponderance of the evidence is that Romney continues to lead here; you’ll get no argument from me there. But I’m curious on what basis you conclude a Romney surge is happening in Arizona. Where do you see that?
February 25th, 2012 at 11:12 am
24
If you want to see all the recent polls, what dates they were taken, and the numbers, go here.
Rasmussen had Romney up by 13 in Arizona. I believe you’re mixing up the Rasmussen Michigan poll with the Rasmussen Arizona poll.
ARG is an outlier in Arizona, considering that the AZ and CNN polls were both taken from Feb 17 to 20, while the other polls having Romney up by double digits were all taken starting the 19th or later.
Rasmussen is the only post-debate AZ poll, and it has Mitt by 13.
http://www.realclearpolitics.com/epolls/2012/president/az/arizona_republican_presidential_primary-1622.html
February 25th, 2012 at 11:13 am
I meant to say the PPP and CNN polls were taken from Feb 17 to 20.
February 25th, 2012 at 11:18 am
#25:
ARG was conducted between Feb 23-24, clearly after voters had had a chance to digest the debate while Rasmussen by his own admission conducted his poll literally within minutes after the debate ended.
Knee-jerk reaction vs. more considered thought + knowing the Rasmussen results and realizing if Romney was to be stopped it had to happen ASAP. I’m talking about the Gingrich supporters here. Again Santorum went up 4 points and Newt went down 4 points. I don’t think that is a coincidence.
Finally until the ARG poll, Newt was consistently polling in the 15-16% range in Arizona. Could this poll be the first sign his supporters are starting to consider moving over the Santorum in big numbers? We’ll soon find out.
February 25th, 2012 at 11:19 am
Romney is going to win by at least 10 in Arizona, and at least 5 in Michigan.
February 25th, 2012 at 11:21 am
27
The ARG poll is not even worth considering, Bob. It has Romney leading by 21 among women and losing by 11 among men. Nothing like that has occurred anywhere this cycle. Nothing even close.
Furthermore, ARG has Santorum gaining substantially among Tea Partiers after the debate where he was called out for endorsing Arlen Specbamacare and voting for things he didn’t believe in because of his “team” of Washington insiders.
You’re wrong, and you will see on Tuesday. Be here.
February 25th, 2012 at 12:00 pm
25. No, no, I assure you I don’t have the Rasmussen Arizona and Michigan polls confused.
Here’s the latest Rasmussen Arizona survey, taken 2/23, put through the PSR filter:
Rasmussen 2/23 +/-4
Romney 42, 38-46
Santorum 29, 25-33
Gingrich 16, 12-20
Paul 8, 4-12
Undecided 4, 0-8
And here’s the latest ARG survey, 2/23-24, put through the same filter:
ARG 2/23-24 +/-4
Romney 39, 35-43
Santorum 35, 31-39
Gingrich 11, 7-15
Paul 9, 5-13
Undecided 6, 2-10
Look at their PSRs. Rasmussen has Romney at 38-46%; ARG has Romney at 35-43%. Rasmussen has Santorum at 25-33% while ARG has him at 31-39%. In other words, one can derive a CPSR from the two surveys because their PSRs overlap/match/agree, or, as Teemu picturesquely puts it, they “touch.”
These two surveys do not disagree; they agree.
As for the PPP and CNN polls, I think it’s important to keep in mind that those two surveys were taken almost the same time as the NBC/Marist survey: PPP 2/17-19, CNN 2/17-20, NBC/Marist 2/19-20. And those three surveys all AGREE. I would add that the Rasmussen survey from 2/16 also agrees with those other three surveys. The CPSR picture for those four surveys is as follows, with trend indicators added based on the Rasmussen survey taken on the 1st:
ARIZONA Composite 2/16-20 Final
1. Romney 39.5-40.5 -3.5-12.5
2. Santorum 28.2-30.5 +11.2-21.5
3. Gingrich 13.5-19 -1-14.5
4. DK 4-11.5
4. Paul 4.2-6.5
So I guess I would agree with you that there appears to have been a break/change, but it seems to have happened just after the 20th, not before. The numbers showing that change/break first show up in the Rasmussen 2/23 survey, not in the NBC/Marist survey.
February 25th, 2012 at 12:17 pm
30
Have you ever seen a bell curve, and looked at where the confidence intervals are placed?
In the case of a poll with a 4% margin of error, 95% of the area of the bell curve is within 4% of the result.
But there is not equal statistical probability that the results are somewhere in that -4 to +4 range. The areas under the curve are orders of magnitude smaller as you approach the end of the confidence interval. The vast majority of the likelihood of a poll is within just 1 or 2% from the polled result.
As such, the likelihood that Rasmussen and ARG “touch” is so small as to be considered nonexistent.
The variance between the two polls is not based on statistical variance. It’s based on poor sampling and even number fudging. ARG is the worst pollster out there, Rasmussen is one of the best.
You need to consider far more than this “PSR” baloney. You need to consider how the race feels, and which polls are more reliable.
Again, ARG and Rasmussen do not “agree.” The probability that they do is so small, that we can say with confidence that they do not.
February 25th, 2012 at 12:29 pm
Also, I will explain MoEs for a second:
The declaration that the result is 95% probable to be within 4% of the polling number is misleading.
What the MoE represents, statistically, is the expected variance WITHIN THE SAMPLE.
For example, if you wanted to know if people enjoyed bacon, and you went to Israel to take the poll, and pretended your poll represented the whole world, your margin of error accounts for the error within your SAMPLE, not the expected error for the true results.
So if in Israel you found that 5% of people enjoyed bacon, and told observers that 5% of the world enjoys bacon, your MoE is still representing the variance within your SAMPLE, not the variance from the truth.
February 25th, 2012 at 12:37 pm
If the MoE represented the expected error from the final RESULT, then the odds the final result in the voting is 8% off is something like 0.005. Yet, it happens ALL the time. That bolsters my comments in #32. The MoE is the error within a sample. If your sample is bad, your error can be 0%, but your final result will be incorrect for the vote count.
February 25th, 2012 at 12:53 pm
30. Let’s say I accept all of that, accept ARG is an outlier, etc. etc.
Furthermore, let’s narrow the PSR as per your stricture that its range should not be the equal of the full margin of error but rather should only extend 2% above or below the central number.
Using that as our standard for PSR, here’s what results (and I have no idea how this is going to come out, since I’m doing this as I’m writing it).
Three previous surveys then agree with each other, Rasmussen 2/16, PPP 2/17-19 and CNN 2/17-20, with following composite result:
ARIZONA Composite 2/16-20 Final
Romney 37-38
Santorum 31-33
Gingrich 16-17
Paul 7-8
Undecided 6-9
On the other hand, two more recent surveys, NBC/Marist 2/19-20 and Rasmussen 2/23, do NOT agree with those three surveys, but they do agree with each other. The composite of those two more recent surveys, plus trends indicated by comparing the result with the composite above, results in the following:
ARIZONA Composite 2/19-23 Final
1. Romney 41-44 +3-7
2. Santorum 27-29 -2-6
3. Gingrich 14-18
4. Paul 9-10 +1-3
5. DK 2-5 -1-7
In other words, using your tighter strainer of no more than 2% above or below the central number, results in a picture that shows Romney climbing since the 20th and Santorum declining within that same period. Interestingly, it also shows Paul climbing and Undecided declining.
I see your point, but it involves not only accepting, at face value, that ARG is “no good”, but also that the other surveys are significantly more accurate, i.e. with tighter margins of error, than I was prepared to accept.
Yes, you were right about Florida (and I have the scars to prove it), but I’m still not convinced. Those two assumptions that I enumerated in the above paragraph strike me both as a bit of a stretch.
February 25th, 2012 at 1:26 pm
34
There is no numerical cutoff point in the MoE. The poll could be within 0% of the truth, it could be within 1%, 2%, 3%, etc. But the farther you go from the horserace numbers, the probability that you are looking at the truth is significantly decreased.
You don’t have to buy that ARG is crap or that the truth is within 2% of the horserace numbers. You just have to understand the difference between samples, and the difference between a sample and the actual vote count.
Take a look at the bell curve. The area under any particular range represents the probability that the truth is within that range.
http://www.google.com/imgres?hl=en&biw=1366&bih=677&gbv=2&tbm=isch&tbnid=X_4qre28C1iStM:&imgrefurl=http://cpmc.coriell.org/Sections/Medical/Risk_mp.aspx%3FPgId%3D97&docid=BL1EI-Hg1Qt1dM&imgurl=http://cpmc.coriell.org/images/c_interval.gif&w=352&h=273&ei=ZzVJT8b6EeLX0QGU94iCDA&zoom=1&iact=hc&vpx=1070&vpy=184&dur=159&hovh=198&hovw=255&tx=196&ty=152&sig=103099642411370841051&page=1&tbnh=135&tbnw=176&start=0&ndsp=18&ved=1t:429,r:5,s:0
February 25th, 2012 at 1:30 pm
So you can see, that if you take two of those bell curves (one representing Rasmussen, one representing ARG), and placed them along a horizontal axis, with, say Santorum’s number as the x-axis set. Put Rasmussen’s bell curve at 29%, and ARG’s bell curve at 35%. The bell curves would indeed overlap, but the area of the intersection between the two bell curves would be so small compared to the overall area of the two bell curves combined, that the overlap would be rendered negligible.
The truth is that ARG and Rasmussen are in parallel universes, and deciding which one is the truth comes down to feel. How does the race feel? What is going on in the race? How is the coverage going? What does momentum suggest? And so forth.
February 25th, 2012 at 3:52 pm
Two weeks ago, we were talking about which states Romney and Santorum would be dividing among themselves on Super Tuesday. Now we’re talking about which states Santorum and Newt will be fighting over, with Romney easily taking the rest.
Newt probably takes Georgia, but I could see an Iowa situation where Romney goes for the knockout punch if it’s close. If Newt wins Georgia, he’s finished.
I’ve been wrong countless times, but if Romney wins both Michigan and AZ with convincing victories and takes the bulk of states on Super Tuesday, GOP voters are going to start focusing on the General with Romney vs Obama. Whether Sanotrum and Gingrich want to face reality or not is a non-factor.
February 25th, 2012 at 5:26 pm
36. Would it not be possible to quantify it more than that?
Am I interpreting correctly that the center of that bell curve to which you reference me represents a statement to the effect that there is a 95% chance that the true number resides in the center of that bell curve?
If that is the case, I’d love to have a system to extrapolate where the slope of the bell curve hits the 50% mark on either side of its center. Perhaps that could be used as the floor and ceiling of the PSR, rather than the margin of error, to which you objected because the margin of error encompasses too large a span of the bell curve, if I understood you correctly.
February 25th, 2012 at 11:51 pm
No, it’s a 95% chance that the true number resides somewhere within the shaded area of the bell curve.
February 25th, 2012 at 11:52 pm
And there is a way to quantify where the bell curve hits the 50% mark, but I don’t remember the formula for finding that out. I’m sure you could find the formula on YouTube or Wikipedia or something.
February 26th, 2012 at 9:28 am
[...] Source: http://race42012.com/2012/02/25/poll-watch-landmark-communicationsrosetta-stone-georgia-2012-republi… [...]
February 26th, 2012 at 8:26 pm
black America, African Americaans, Historically Black Colleges and Universities, black fraternities, Black clubs, Morning star, black churches, black communities,…
Race42012.com | 2012 GOP News, Polls and Blogs for Republicans…