The Phantom's Polling Analysis Thread

[BluePhantom]

An excellent thread Phantom, and I hope we can keep it active all the way to election.

I think polling is instructive to the candidates at this time as they let them know how this or that rhetoric or tactic is playing in Peoria, and the polls do have a bearing on those tactics and the jargon we will hear in the campaign.

Polling in April and May isn't that useful as to who is likely to win in November as that will largely be determined on a number of things including the economy, whether any major scandal surfaces, or whether either candidate makes a major tactical blunder.

I agree completely. I think the actual spreads right now are only significant in regards to establishing a starting point. My personal opinion is that Romney will close anywhere from 5%-8% between now and November for a variety of reasons (choosing a running mate, exposure in the debates, healing within the GOP, etc) so what I am really looking at are two things: a) what is the spread to start with? Is it under 8%? If so, it's in play. b) are trends being established that are showing that Romney is closing that gap or is Obama successfully defending it? Are those trends supporting that 5%-8% closing projection?

Now that's just how I am looking at it personally, which I am of course happy to discuss. As PP noted below I really want to keep the "official reports" statistically accurate so we are discussing these questions and the points you pose from a sound foundation.

 

[skookerasbil]

Phantom........just check the Rasmussen polls. They are pretty spot on. Back in the early fall of 2008, the left guys in here couldnt post up the Rasmussen polls fast enough, in fact, they were falling all over themselves on a daily basis to do so. And at election time, they were deadly accurate. Rasmussen of course, polls likely voters.............most other polls are chasing the hopelessly duped who dont even know what a voting booth is.

Now, of course, the k00ks hate Rasmussen since Obama tanked in the summer of 2009.

 

[PLYMCO_PILGRIM]

An excellent thread Phantom, and I hope we can keep it active all the way to election.

I think polling is instructive to the candidates at this time as they let them know how this or that rhetoric or tactic is playing in Peoria, and the polls do have a bearing on those tactics and the jargon we will hear in the campaign.

Polling in April and May isn't that useful as to who is likely to win in November as that will largely be determined on a number of things including the economy, whether any major scandal surfaces, or whether either candidate makes a major tactical blunder.

It is too early for the polls to mean much for the election but they do show how the general populace feels about the canidates in light of the different things that are active in the news.

 

[skookerasbil]

I do find it amazing that Obama's poll numbers are stilll this high. Clearly, the landscape has changed a bit as compared to the last few decades of the modern era. In the past, with the economy being such a joke, no way a president would poll in the mid to high forties. Clearly, the sharp increase in handouts over the past few years has won over more people.

Still.........no worries here. Every swinging dick out there who is against an Obama reelection will be voting and the women are just as passionate. Zero % chance the passion is the same as it was in 2008 on the side of the hopelessly duped........in fact, nobody has heard a thing about "Hope and Change" for a long, long time now. All the dolts who thought Obama would be payng for their new kitchens and their car = fAiL. They'll be sitting on their hands come election day..........still wondering what the fuck happened?

Nobody knows jack about Romney and he's now tied. The key swing states like Florida and Ohio will go red, as will Virginia. Thats 60 electoral votes right there s0ns. Its all good!!

 

[BluePhantom]

I do find it amazing that Obama's poll numbers are stilll this high. Clearly, the landscape has changed a bit as compared to the last few decades of the modern era. In the past, with the economy being such a joke, no way a president would poll in the mid to high forties. Clearly, the sharp increase in handouts over the past few years has won over more people.

I am not surprised at all. Obama's policies are unpopular but Obama himself is very popular. People genuinely like him as a person and that will strongly influence a person's vote. Also keep in mind that, again, almost all of these polls are RV polls and by the very nature of an RV poll it will inflate the Democratic candidate's numbers. In 2010, Nate Silver estimated this effect to be about 6%. Seems a bit high to me...I'd say it's probably more like 4% - 4.5% but regardless it's probably somewhere between 4% and 6% so let's split the difference at 5%.

Well what that means is a couple things: a) when you look at an RV poll or consider an average of RV polls, if you take the spread and reduce Obama's share by 5% you are probably a little closer to reality because what you are essentially doing is taking into account that many of those registered voters will not actually go vote...and because registered Republicans and Independents vote at a higher percentage than registered Democrats it brings the numbers a little closer to what would probably happen if the election was held today. There are problems doing that of course and it doesn't work quite that way so I can't endorse looking at it that way from a strict statistical perspective but it's a fair argument to make when you are speculating.

b) Around September all these polling agencies will stop releasing RV data and will start releasing LV data exclusively. This will create the appearance of a Romney surge of between 4% - 6% on average right around that time. In reality all that is happening is that the polling agencies are suddenly going to change their methodology and the picture will become more accurate. But that's important because even though that surge will be an illusion it will grab the public's attention and you may see a bandwagon effect that enhances it and creates a real Romney surge afterwards.

Of course all this is irrelevant if the economy turns around and we start to see real recovery. If that happens any gains Romney has made will dry up and the trends will turn in Obama's favor.

 

[BluePhantom]

Well it's the last day of the polling week and I am eagerly awaiting the final polls to be posted on RCP for this week's analysis. There is much more data this week so there will be some good information to dive into. The report will come tomorrow as I want to give the full day today for any stragglers to come in before I put together the report.

Anyhow I noticed something that struck me as odd today and it's prompted me to introduce a new report which I will do....oh perhaps every other month or so. I saw a poll today by a specific professional polling agency that was way out of line with the information provided by the majority. My first thought was that it was simply an outlier but when I looked at the name of the company I realized that it was like the third time in the last month that they had released a poll that really seemed to be out in left field. There are two possibilities: they are correct and everyone else is wrong, or there is something in their polling method that is consistently providing data heavily weighted to one side. So what I have decided to do is to....uh.....essentially "normalize" each agency's data in an effort to determine the "house bias" for each firm based on available current data (ignoring historical information from past elections, etc) and release that information in a separate report. Not sure what I will call it yet...."Agency Average Deviation Report" comes to mind.

I don't view this as an exercise intended to discredit this firm or that one, or create favorable impressions of this firm vs. that firm, etc; at least not for the purposes of "official reporting" which will remain the same regardless of what this investigation reveals....because there are several reasons why such a trend in a given firm might exist and not all of them are necessarily "evil plots". It's primarily to satisfy my personal curiosity and perhaps provide some interesting "food for thought"...perhaps even some elements for consideration that can be used in other debates on other threads.

First I have to finish this week's polling, demographics, and trends reports so I doubt I will get to it until near the end of the week...but I am really curious to see what the results will be.

 

[Foxfyre]

And we might as well go ahead and acknowledge the other component that is ever present in analysis of Barack Obama. It is REQUIRED that he be popular personally because he is designated black. So you call somebody for a list of political questions and that person goes right down the line expressing disapproval of almost all of Obama's policy initiatives and proposals. But when they get to the question re their approval of Obama personally, they might think he is a schmuck, a Marxist, an incompetent, or a villain, but they will say they have nothing against him personally. They don't want to be accused of disliking him because he is black.

So it goes into the approval column. I don't recall any other scenario, except for Jimmy Carter, in which people who disapproved of a President's policies gave the President a high approval rating personally. But of course my perceptions could be flawed here as I'm not doing any polling.

As we start getting down to the nitty gritty where polling data does start indicating who people are likely to vote for, it will be interesting to see if that changes. I honestly don't know how many people genuinely like and approve of Barack Obama.

 

[BluePhantom]

And we might as well go ahead and acknowledge the other component that is ever present in analysis of Barack Obama. It is REQUIRED that he be popular personally because he is designated black. So you call somebody for a list of political questions and that person goes right down the line expressing disapproval of almost all of Obama's policy initiatives and proposals. But when they get to the question re their approval of Obama personally, they might think he is a schmuck, a Marxist, an incompetent, or a villain, but they will say they have nothing against him personally. They don't want to be accused of disliking him because he is black.

So it goes into the approval column. I don't recall any other scenario, except for Jimmy Carter, in which people who disapproved of a President's policies gave the President a high approval rating personally. But of course my perceptions could be flawed here as I'm not doing any polling.

As we start getting down to the nitty gritty where polling data does start indicating who people are likely to vote for, it will be interesting to see if that changes. I honestly don't know how many people genuinely like and approve of Barack Obama.

I think there is probably some truth to that and I have seen some data that suggests that Obama does not perform nearly as well in automated polls as he does in direct human contact polls. The speculation of course is that when someone is taking the poll and talking to another human being on the other end of the line, they are hesitant to oppose Obama for fear of being considered a racist (whether consciously or sub-consciously) but in automated polls they don't have that fear and respond more freely. I think there is probably some merit to that as there have been some elections historically where the polls said one thing about a black candidate across the board but when the election results came in they were completely different. But tracking it statistically would be very difficult because there are so many things that can bias a poll. I mean we may be able to identify an average difference in support between automated and human contact polls, but proving that that's the reason for that difference would be quite an undertaking...frankly probably impossible.

 

[Dr.Traveler]

as of todays poll giving Obama 49% ? really, half of America is loving the 4.00 gas and high food prices and knowing their kids are already $50,000 in debt?

It's more a commentary on Romney. He struggled to carry 40% in his own primary process, he'll probably struggle to carry 51% at his best in this process. But we'll see.

You can win a Congressional seat by not being the incumbent, but you can't win the White House that way.

 

[Dr.Traveler]

Obama Grand Total: 247
Romney Grand Total:239


Too Close to Call

Colorado (9), Iowa (6), Nevada (6), Ohio (18), Virginia (13)

Total TCTC: 52



As a side note: for practical consideration a tie will go to Romney as the tie is broken by the House of Representatives with each state getting one vote. In that scenario Obama would likely get a minimum of 21 votes and a maximum of 24. Romney would get a minimum of 26 and a maximum of 29.

Let's take a look at this for a second. The common wisdom is that the election will come down to six states (Nevada, Ohio, Florida, Virginia, Iowa, and New Hampshire), but if we accept the EV breakdown I have, which seems reasonable enough, then at the end of the day Obama would need 23 electoral votes to win and Romney would need 31 from those states I listed as "too close to call". Now let's give Obama Colorado, Nevada, and Iowa. That's only 21 EVs for a total of 268. That means Obama MUST take either Ohio or Virginia. Romney can give up CO, NV, and IA and reach 270 if he takes both Ohio and Virginia. Ohio is trending toward Romney right now and Obama's lead in Virginia is within the MOE. Obama's lead in Ohio is also within the MOE. In other words this election will really come down to Ohio and probably Virginia.

This is really good news for Romney supporters and it should absolutely terrify Obama supporters.

I've been looking at the electoral math too, and agree. The electoral math seems to lean towards a Romney win. Especially since all of the calculations I've seen include Pennsylvania in the Obama column, and I just can't see that being even close to a safe bet this time. And if Pennsylvania is in play, a lot of other states in the Obama column could be.

There's an analysis at Electoral-Vote.com that follows a similar line. They're a little more optimistic for Obama though.

I do think it's hilarious that there's still the possibility of a tie, as it indicates just how weak both candidates are. As an incumbent with fairly high personal likeability scores, Obama should be walking away with this. As the challenger in a tough economy following up on a strong party surge, Romney should be walking away with this. You have two candidates with incredibly strong indicators that are both struggling to get to 51%.

 

[Foxfyre]

And we might as well go ahead and acknowledge the other component that is ever present in analysis of Barack Obama. It is REQUIRED that he be popular personally because he is designated black. So you call somebody for a list of political questions and that person goes right down the line expressing disapproval of almost all of Obama's policy initiatives and proposals. But when they get to the question re their approval of Obama personally, they might think he is a schmuck, a Marxist, an incompetent, or a villain, but they will say they have nothing against him personally. They don't want to be accused of disliking him because he is black.

So it goes into the approval column. I don't recall any other scenario, except for Jimmy Carter, in which people who disapproved of a President's policies gave the President a high approval rating personally. But of course my perceptions could be flawed here as I'm not doing any polling.

As we start getting down to the nitty gritty where polling data does start indicating who people are likely to vote for, it will be interesting to see if that changes. I honestly don't know how many people genuinely like and approve of Barack Obama.

I think there is probably some truth to that and I have seen some data that suggests that Obama does not perform nearly as well in automated polls as he does in direct human contact polls. The speculation of course is that when someone is taking the poll and talking to another human being on the other end of the line, they are hesitant to oppose Obama for fear of being considered a racist (whether consciously or sub-consciously) but in automated polls they don't have that fear and respond more freely. I think there is probably some merit to that as there have been some elections historically where the polls said one thing about a black candidate across the board but when the election results came in they were completely different. But tracking it statistically would be very difficult because there are so many things that can bias a poll. I mean we may be able to identify an average difference in support between automated and human contact polls, but proving that that's the reason for that difference would be quite an undertaking...frankly probably impossible.

That is true. I participate sometimes in automated on line polls for Zogby. I do respond differently emotionally to an automated poll while I at least pretend to be thoughtful when speaking with a person asking the questions. And that probably does cause me to answer some questions to a greater extreme than I would if it was a person calling me.

But agreed that tracking it statistically would be impossible. Or at least would be beyond my polling or statistics skills.

 

[BluePhantom]

I've been looking at the electoral math too, and agree. The electoral math seems to lean towards a Romney win. Especially since all of the calculations I've seen include Pennsylvania in the Obama column, and I just can't see that being even close to a safe bet this time. And if Pennsylvania is in play, a lot of other states in the Obama column could be.

Since Santorum dropped out we only have one poll of Pennsylvania. It was a Quinnipiac poll and they are my 2nd ranked pollster in terms of accuracy and they showed an 8 point Obama lead at the end of last week. They are a very reliable agency but even the best can fall victim to the dreaded outlier so I would like so see more polling to serve as a basis for comparison. Historically, even though everyone says PA is a swing state, it hasn't gone to a Republican candidate since 1988. I think after 24 years it's time to say they are a pretty reliable blue state and that Romney is going to have to really do something impressive to steal it. At this point I don't see it happening because Philadelphia (especially) and Pittsburgh (to a lesser extent) completely overpower the rest of the state. I'd say Romney taking PA is possible but I wouldn't bet the family jewels on it right now.

 

[BluePhantom]

And we might as well go ahead and acknowledge the other component that is ever present in analysis of Barack Obama. It is REQUIRED that he be popular personally because he is designated black. So you call somebody for a list of political questions and that person goes right down the line expressing disapproval of almost all of Obama's policy initiatives and proposals. But when they get to the question re their approval of Obama personally, they might think he is a schmuck, a Marxist, an incompetent, or a villain, but they will say they have nothing against him personally. They don't want to be accused of disliking him because he is black.

So it goes into the approval column. I don't recall any other scenario, except for Jimmy Carter, in which people who disapproved of a President's policies gave the President a high approval rating personally. But of course my perceptions could be flawed here as I'm not doing any polling.

As we start getting down to the nitty gritty where polling data does start indicating who people are likely to vote for, it will be interesting to see if that changes. I honestly don't know how many people genuinely like and approve of Barack Obama.

I think there is probably some truth to that and I have seen some data that suggests that Obama does not perform nearly as well in automated polls as he does in direct human contact polls. The speculation of course is that when someone is taking the poll and talking to another human being on the other end of the line, they are hesitant to oppose Obama for fear of being considered a racist (whether consciously or sub-consciously) but in automated polls they don't have that fear and respond more freely. I think there is probably some merit to that as there have been some elections historically where the polls said one thing about a black candidate across the board but when the election results came in they were completely different. But tracking it statistically would be very difficult because there are so many things that can bias a poll. I mean we may be able to identify an average difference in support between automated and human contact polls, but proving that that's the reason for that difference would be quite an undertaking...frankly probably impossible.

That is true. I participate sometimes in automated on line polls for Zogby. I do respond differently emotionally to an automated poll while I at least pretend to be thoughtful when speaking with a person asking the questions. And that probably does cause me to answer some questions to a greater extreme than I would if it was a person calling me.

But agreed that tracking it statistically would be impossible. Or at least would be beyond my polling or statistics skills.

Mine too...especially given the data that the firms release. So often their crosstabs are so lacking (or flat out non-existent) that we are forced to extrapolate on something as basic as the political ideologies of the survey sample. To track the effect we are discussing to even the most remote level of accuracy would be impossible without far far deeper data....and even THEN it would still be tough because as I said....how do you prove that the respondent is answering a given way because of that effect? I think the best we can do is to file the effect away in our brains and in a situation where there's a point spread of less than 1%...now we might want to take it into consideration a BIT.

 

[BluePhantom]

Weekly Breakdown for 5/8/12

A far better week in regards to the number of polls released. There were three professional polls and three media polls released for the general election as well as several state polls for analysis. As always I am frustrated with the lack of extensive demographic data but there is certainly more than last week so we should take it and run, I suppose.

As always there were those polls that seemed to support the current trends and those that raise eyebrows and seem to be outside the majority. As such we have some very interesting results and supporters of both Romney and Obama have something to cheer about and something to cringe over. Without question there will be plenty to look at and debate. This week’s bottom line is that Romney is picking up speed on a national level. After four weeks of analysis his trendline on my spreadsheet in now ascending while Obama’s is descending. Projected through November, if current trends continue, Romney appears on pace for a minimum 3% victory in the popular vote. If that trend continues that gap will grow over future weeks. That’s the good news for Romney. The good news for Obama is that in a couple swing states the data suggests he is tightening his grip over Romney. This creates a little confusion. Why would Romney be surging nationally, but suddenly declining in these swing states? There could be a lot of reasons, and I certainly have my theories, but I will leave those for future discussion as the purpose of this report is not to inject opinion. As usual all links can be found on RCP.

With that said let's look at the statistics:

5/1/12 - 5/8/12

Base Average
Professional Polls
Rasmussen (LV - tracking): Obama 44, Romney 49
Gallup (RV - tracking): Obama 44, Romney 47
Democracy Corps (D) (LV): Obama 47, Romney 47

Professional Average: Obama 45.00, Romney 47.67; Romney +2.67

Media Polls
Reuters/Ipsos (RV): Obama 49, Romney 42
Politico/GWU/Battleground (LV): Obama 47, Romney 48
IBD/CSM/TIPP (RV): Obama 46, Romney 43


Media Average: Obama 47.33, Romney 44.33; Obama +3

Overall Base Average: Obama 46.17, Romney 46.00; Obama +0.17


Adjusted Average
Professional Polls
Rasmussen (LV - tracking): Obama 660, Romney 735
Gallup (RV - tracking): Obama 968, Romney 1034
Democracy Corps (D) (LV): Obama 470, Romney 470

Total: Obama 2098, Romney 2239

Professional Average: Obama 44.64, Romney 47.64; Romney +3.00

Media Polls
Reuters/Ipsos (RV): Obama 469.91, Romney 402.78
Politico/GWU/Battleground (LV): Obama 470, Romney 480
IBD/CSM/TIPP (RV): Obama 393.76, Romney 368.08

Total: Obama 1333.70, Romney 1250.90

Media Average: Obama 47.38, Romney 44.44; Obama +2.94
Overall Adjusted Average: Obama 45.66, Romney 46.44; Romney +0.77


Two Week Rolling Average

Base Average
Professional Polls

Obama 45.50, Romney 46.83; Romney +1.33

Media Polls

Obama 46.92, Romney 43.42; Obama +3.50

Combined Average

Obama 46.21, Romney 45.13; Obama +1.08

Adjusted Average
Professional Polls

Obama 45.24, Romney 46.92; Romney +1.68

Media Polls

Obama 47.03, Romney 43.59; Obama +3.45

Combined Average

Obama 45.88, Romney 45.72; Obama +0.17

In state polls there was some more bad news for Romney continuing over from last week. A PPP(D) RV poll showed Obama with a 51/41 lead in the state of Iowa. There have been no other polls in Iowa since Santorum dropped out of the race so there is no other data to support or refute those results.

In Ohio a PPP (D) RV poll showed Obama with a 50/43 lead over Romney. This is quite surprising as the data from the last two weeks by Rasmussen and Quinnipiac showed a much tighter race and a trend toward Romney. There could be several reasons why the data shows Obama’s lead in Ohio jumped from two points to seven in the last week, but as it is a direct reversal of previous trends and a pretty dramatic jump it’s worth some raised eyebrows. Is the poll accurate and the trend has shifted or is the poll an outlier? Time will tell.

In Virginia the ABC/Washington Post RV poll showed a 51/44 Obama advantage. Individually this supports the PPP (D) poll of Virginia from a week ago but stands at odds with Rasmussen and Purple Strategies who show a far closer race. Because the ABC/WaPo poll was the only poll of Virginia this week the data shows a dramatic increasing trend for Obama in Virginia, but it remains to be seen who is correct as we have two polls showing one thing and supporting each other and two poll showing something completely different and supporting each other. Again, time will tell.

 

[BluePhantom]

A little personal commentary for a moment (this is where I get to throw in some opinion ). I am really looking closely at those PPP polls. Something with them doesn't appear to be right with me. As I go back and look at the data over the last four weeks, with only a couple exceptions they have been really at odds with the majority of the polls and demonstrating a very strong Obama bias. In some cases it's been 5-6 points above the average of what everyone else is showing. I wouldn't worry so much about it if it was only one or two polls that showed this kind of lean because I could write it off as an outlier which every polling organization will get now and again; but there are too many of them to write off as simple "freak accidents".

I have to argue with myself a little because I say "well...is it that they are at odds with the average so badly or is it because you personally don't like what they suggest?" Well, I have to concede that I don't like what they say but it's really the former that is bothering me. I have the same issues with media polls because they show an indisputable bias of 5-6 points on average on top of the natural RV bias of roughly 5 points. But it's data and I include them all in the calculations whether I like them or not. The point is that PPP is suddenly showing results similar to media polls and is really at odds with the professional polling organizations on a consistent basis.

I also say to myself: "well Rasmussen is showing an average 2.71 lean toward Romney against the average. You don't seem to have as much problem with that." Yeah....but Rasmussen does LV polls compared to everyone else's RV polls so that's really not surprising. That makes perfect sense because an LV poll eliminates the natural biases that an RV poll has.

So why is PPP suddenly so far out of the mainstream all the time? Is it that they are run and funded by the Democratic Party? Perhaps but historically their bias has never been this dramatic. I mean it's always been there a little bit but historically they have shown a reasonable degree of reliability and frankly in the 2006 elections, if I recall correctly, they were actually showing a very slight bias toward the GOP. So I am really at a loss to explain this. Hopefully, it's just a series of freak events (and from my personal point of view, God forbid they are right and everyone else is wrong ) and in the future they will produce results that are more commonly supporting other data.

 

[BluePhantom]

So as I am going through the demographics and breaking them down I have noticed something amusing. The Politico poll is listed as an LV poll. Their basis for a "likely voter" is asking "are you likely to vote in the upcoming election?" Classic. In other words it's an RV poll masquerading as an LV poll. You know shit like this just annoys the hell out of me. Then in question 2 the phrasing is:

"If the election for President were being held today, and you had to make a choice, would you be voting for -
- (ROTATE) --
• the Republican candidate or
• President Barack Obama, the Democratic candidate?"

They don't even mention Romney by name but refer to Obama intentionally as the President. Despite this Romney leads in that poll by a point, but it's precisely that kind of crap which is why media polls are held in such low esteem.

http://images.politico.com/global/2012/05/bg_47_questionnaire.html

 

[Foxfyre]

A little personal commentary for a moment (this is where I get to throw in some opinion ). I am really looking closely at those PPP polls. Something with them doesn't appear to be right with me. As I go back and look at the data over the last four weeks, with only a couple exceptions they have been really at odds with the majority of the polls and demonstrating a very strong Obama bias. In some cases it's been 5-6 points above the average of what everyone else is showing. I wouldn't worry so much about it if it was only one or two polls that showed this kind of lean because I could write it off as an outlier which every polling organization will get now and again; but there are too many of them to write off as simple "freak accidents".

I have to argue with myself a little because I say "well...is it that they are at odds with the average so badly or is it because you personally don't like what they suggest?" Well, I have to concede that I don't like what they say but it's really the former that is bothering me. I have the same issues with media polls because they show an indisputable bias of 5-6 points on average on top of the natural RV bias of roughly 5 points. But it's data and I include them all in the calculations whether I like them or not. The point is that PPP is suddenly showing results similar to media polls and is really at odds with the professional polling organizations on a consistent basis.

I also say to myself: "well Rasmussen is showing an average 2.71 lean toward Romney against the average. You don't seem to have as much problem with that." Yeah....but Rasmussen does LV polls compared to everyone else's RV polls so that's really not surprising. That makes perfect sense because an LV poll eliminates the natural biases that an RV poll has.

So why is PPP suddenly so far out of the mainstream all the time? Is it that they are run and funded by the Democratic Party? Perhaps but historically their bias has never been this dramatic. I mean it's always been there a little bit but historically they have shown a reasonable degree of reliability and frankly in the 2006 elections, if I recall correctly, they were actually showing a very slight bias toward the GOP. So I am really at a loss to explain this. Hopefully, it's just a series of freak events (and from my personal point of view, God forbid they are right and everyone else is wrong ) and in the future they will produce results that are more commonly supporting other data.

It is worse than that in New Mexico in advance of our Primary coming up next month. We are one of the battleground states of course meaning that we get more national ads and more attention from the candidates than do the states in which it is presumed the election is already decided.

So for weeks now we have been inundated with Democratic ads as how the Democrat will protect the people's social security, jobs, student loan rates, etc. etc. etc. and how the Tea Party Republicans--specific term being used in almost everybody's campaign--want to take all those away and give more tax breaks to the rich, yadda yadda.

So what phrase shows up in some of the telephone polls we have received at the house? Will you vote for the Democrat or the Tea Party Republican? So blatant it is embarrassing.

 

[BluePhantom]

It is worse than that in New Mexico in advance of our Primary coming up next month. We are one of the battleground states of course meaning that we get more national ads and more attention from the candidates than do the states in which it is presumed the election is already decided.

So for weeks now we have been inundated with Democratic ads as how the Democrat will protect the people's social security, jobs, student loan rates, etc. etc. etc. and how the Tea Party Republicans--specific term being used in almost everybody's campaign--want to take all those away and give more tax breaks to the rich, yadda yadda.

So what phrase shows up in some of the telephone polls we have received at the house? Will you vote for the Democrat or the Tea Party Republican? So blatant it is embarrassing.

Yeah I am looking at the breakdown of the Democracy Corps (D) poll now. They don't ask the question of Romney vs. Obama until question #39. So they spent 38 questions influencing the respondent with their phrasing of their questions before they pop the big question of who they support. Now it turned out to be a 47/47 tie in that particular poll, so it's not the results I am upset with. It's that a rank amateur structuring his first poll knows better than to do that because the results will be completely biased and unreliable. Just makes me want to scream.

 

[BluePhantom]

5/8/2012 Demographics and Trends Analysis


Well it appears I was in error about having more demographic data this week. Upon closer inspection only the TIPPOnline poll provided any crosstabs that could be evaluated. I am not particularly thrilled about this as they are notoriously inaccurate, but it’s all we have and I promised I would include all data whether I like it or not so into the spreadsheet it goes. For the most part their data is similar enough to what we have seen before that I am not terribly concerned about it, although there were some things that really raised my eyebrows and I have a hard time believing. Support among Hispanic voters for example jumping from a 68/30 Obama advantage to an 80/12 Obama advantage in one week is worth some heavy skepticism. They also provided some data on age that I had to play around with a bit to fit into the groupings we have chosen to focus on. They provided data on the age range of 18-44, for example, where we want to split that into 18-29 and 30-45. I used census bureau data that I provided the link for earlier in this thread and attempted to estimate spreads that would meet their 46/43 breakdown while keeping as close as possible to previous data and the trendlines. It’s probably not right on the nose but it should be somewhere in the ballpark.

I need to stress though that like last week these demographics are based on one poll this week and not a very reliable or respected agency at that, so once again…take the results with a huge grain of salt.

Among women Obama holds a 49/40 advantage. This shows a 2 point drop in support for Romney while Obama’s support held steady (a two point swing to Obama’s side) since the last demographics breakdown last week. The data continues to suggest a decreasing trend among women in regards to support for Obama while Romney’s support has held relatively steady over the last month. As the swing is outside the margin of error, it’s statistically irrelevant and the trends are so slight that projecting toward November according to current trends there would hardly be any movement at all. If these trendlines continue Romney is on pace to pick up maybe a point and a half among women between now and election day.

According to Ideology, this poll shows that last week’s results are supported in regards to Independents moving away from Obama although a) not necessarily toward Romney, and b) perhaps not as dramatically as last week’s data suggested. Romney’s support held steady at 46% while Obama climbed back from last week’s low of 33% to 37%. This is still far lower than the 52% support among Independents that Obama enjoyed when we started this analysis four weeks ago. At a two week average this means the split is probably somewhere around 35/46 in Romney’s advantage. If current trends continue Romney would approach election day with around a 13%-14% advantage over Obama among Independent voters.

According to race this poll showed some interesting results as I mentioned above regarding Hispanic voters. All previous polls that have been evaluated have shown Hispanic support to be roughly 68/30 in favor of Obama. This poll’s 80/12 Obama domination seems to me a little difficult to believe. It’s not that I have a hard time believing there would be an increasing trend in support for Obama among Hispanic voters, but a jump that dramatic leads me to hold the personal opinion that the sample either had a very low population of Hispanic voters which resulted in skewed data for that demographic, or perhaps there was just some freak occurrence where every Hispanic voter they surveyed happened to be an Obama supporter. Whatever the cause, the wide disparity between this poll and previous data and trends leads me to believe that this particular information is probably an outlier. Among white voters there was a three point decline in support for Obama from 40% to 37%. That swing would be on the outer edge of the MOE so perhaps there might be a little actual movement there. We will know more in future weeks if that develops into a trend or not. Romney’s support has held pretty firm at 53%-54% with only minimal movement.

According to age this poll again showed some things which seem a little curious. For the 66+ demographic it shows a 49/42 Obama edge. Again…I find that pretty difficult to buy into as older voters are notoriously more conservative and these results show support for Obama to be six points above the trendline and support for Romney to be eleven points below the trendline. Personally, I don’t buy it but that’s what the data says so that’s what we’re going with. In the 46-65 range Romney’s support has ticked down about two points to 44% while Obama’s has dropped five points to 46%. This would seem to indicate that voters in that age range are reconsidering as both Obama and Romney show a drop in support…that means more undecided voters. In the 30-45 age range I estimate (see notes above) a 42/48 Romney advantage. This would indicate a slight uptick in support for Romney but a five point downgrade in support for Obama. Similarly in the 18-29 bracket if my calculations are correct I estimate Romney’s support holding steady at 37% while Obama’s support dropped five points to 51%. However those specific age ranges actually break down in reality when we take those two groups together as a whole this poll does show declining support for Obama among voters aged 18-45 at about a five point clip. Again that’s one hell of a drop in support for Obama against the trendline so my guess is it’s probably not quite that dramatic but again, that’s what the data we have says so that’s what we’re going with.


In state trends

As previously noted in the earlier report, a PPP (D) poll of Ohio shows a dramatic turnaround from Obama’s declining trend in support among Ohio voters that the previous polls have suggested. As mentioned before (in previous posts and in last week’s trend report) I simply don’t have enough data that I am very confident about to say conclusively that there is an undeniable trend here one way or the other. Either Rasmussen and Quinnipiac are correct or PPP is correct, but someone is completely incorrect. Time will tell which one it is and I will let you make that determination for yourself for now until I can show something statistically when future polls on Ohio are released.

ABC/WaPo polled Virginia and showed a 51/44 Obama advantage. This seven point split is quite a departure from the three previous polls last week which averaged out to a three point Obama advantage. Those three polls were all professional polls and the ABC/WaPo is a media poll and on average media polls tend to show an advantage for Obama by roughly 4%-5% in comparison to the professional pollsters so if you take that into consideration it makes a lot more sense. Since we are not weighting data in these calculations it means statistically there is a dramatic pro-Obama trend in Virginia, although in reality it really suggests that there hasn’t been much movement at all since last week and a two week rolling average would indicate a four point Obama advantage, just outside the MOE.

There was not enough data to look for trends in any other state.


So I must emphasize again that the demographics data for this week was taken from one poll and a poll that I don’t particularly trust. Still the data was calculated the same as all other data and the numbers are being reported according to the accurate results of those calculations.

In the end the data is particularly confusing when compared to the overall information from this week’s overall numbers. The trend of an increase in Romney support is clear overall and is supported by multiple polls, yet one or two individual polling agencies have produced some data in specific areas that completely contradict those overall trends. In other words there is a lot of room for interpretation (and argument) about this week’s information.

 

[Foxfyre]

When one of Rasmussen's running polls shows a sharp unexplained deviance from recent trends, he always notes that along with the warning that it is too early to determine whether the new data is a trend or whether it is just 'statisitical noise', i.e. a temporary anomaly that won't hold up. TIPP apparently isn't aware of such things or doens't think it is important enough to mention.