The Phantom's Polling Analysis Thread

[swizzlee]

Yup, I'm a numbers worker, too, but I've been involved in research for over 15 years where other factors also must carry weight in the final analysis. Of course we could all just sit back with a 6-pack and wait til November............

I think the Supremes decision on healthcare will have an impact on voter thinking.

I think the issue with contraceptives and the Catholic church will continue to have an impact, most particularly since the 43 lawsuits have been filed. It was also said that the church plans to organize protests in June. More impact.

Although news has been relatively low key, the outcome of the Brian Terry/DOJ situation could have an impact.

Any or all of the above could change poll numbers............

Playing my own game with the electoral numbers, I currently have Romney with 235, Obama 202.

 

[BluePhantom]

Yup, I'm a numbers worker, too, but I've been involved in research for over 15 years where other factors also must carry weight in the final analysis. Of course we could all just sit back with a 6-pack and wait til November............

man...what fun would that be?

I think the Supremes decision on healthcare will have an impact on voter thinking.

I think the issue with contraceptives and the Catholic church will continue to have an impact, most particularly since the 43 lawsuits have been filed. It was also said that the church plans to organize protests in June. More impact.

Although news has been relatively low key, the outcome of the Brian Terry/DOJ situation could have an impact.

Any or all of the above could change poll numbers............

Playing my own game with the electoral numbers, I currently have Romney with 235, Obama 202.

i think those are all valid points. Right now I have Romney with 239, Obama with 247, and 52 up for grabs

 

[NYcarbineer]

Likely Voter Report for 5/22/12

This is the first LV Report released and we have some good info to delve into. We had two LV polls released this week (Rasmussen and Mason-Dixon) and two polls where enough data was available to convert it from RV format to a generic LV format as an estimate. As previously noted when converting we will be using a standard 32/32/36 D/R/I model to start with, however I will be tracking that breakdown through the election and adjusting that model according to the averages. What this really tells us is how energized the bases are and that can play a significant role and will shift the numbers accordingly. Of course their energy level will fluctuate between now and election day and as a result (once we get more data) the 32/32/36 model will change to reflect that. I am of several minds regarding how to track that as a trend because with multiple floating variables it’s likely that we will see much larger average swings when we adjust the D/R/I model. Ultimately, it won’t much matter because around September everyone will move to LV polls anyhow and so (in theory at least) everything should kind of trend toward the LV stats and mesh together as the averages bump around from week to week.

Regardless, for now we will stick with what we have and if and when it becomes evident that the model needs adjusted I will have made a decision on how to approach the trendlines: do we recalibrate all previous polls to the new model or leave them alone? There are strong arguments for both and it really depends on what we are trying to measure. Anyhow, I did go back to the previous polls and convert them (at least those that were convertible) to LV models as well in order to plot a trendline
Here’s how things break down:

5/16/12 - 5/22/12

Verifiable Professional Polls
Agency Sample %Oba %Rom Spr MOE #Oba #Rom
Rasmussen 1500 44.00% 46.00% R+2.00% 2.28% 660 690
Mason-Dixon 1000 44.00% 47.00% R+3.00% 2.82% 440 470

Converted Professional Polls

TIPP 778 40.44% 42.52% R+2.08% 2.91% 314.62 330.80



Verifiable Media Polls
None

Converted Media Polls
FOXNews 913 40.52% 42.00% R+1.48% 2.68% 369.95 383.46

Base Average
Verified Professional: Obama 44.00, Romney 46.50; Romney +2.50 (MOE: 1.81)
Converted Professional: Obama 40.44, Romney 42.52; Romney +2.08 (MOE: 2.91)
Combined Professional: Obama 42.81, Romney 45.17; Romney +2.36 (MOE: 1.55)

Verified Media: None
Converted Media: Obama 40.52, Romney 42.00; Romney +1.48 (MOE: 2.68)
Combined Media: Obama 40.52, Romney 42.00; Romney +1.48 (MOE: 2.68)

Verified Combined: Obama 44.00, Romney 46.50; Romney +2.50 (MOE: 1.81)
Converted Combined: Obama 40.48, Romney 42.26; Romney +1.78 (MOE: 1.98)
Overall Combined: Obama 42.24, Romney 44.38; Romney +2.14 (MOE: 1.34)

Adjusted Average (weighted according to sample size)
Verified Professional: Obama 44.00, Romney 46.40; Romney +2.40 (MOE: 1.77)
Converted Professional: Obama 40.44, Romney 42.52; Romney +2.08 (MOE: 2.91)
Combined Professional: Obama 43.16, Romney 45.48; Romney +2.32 (MOE: 1.52)

Verified Media: None
Converted Media: Obama 40.52, Romney 42.00; Romney +1.48 (MOE: 2.68)
Combined Media: Obama 40.52, Romney 42.00; Romney +1.48 (MOE: 2.68)

Verified Combined: Obama 44.00, Romney 46.40; Romney +2.40 (MOE: 1.77)
Converted Combined: Obama 40.48, Romney 42.24; Romney +1.76 (MOE: 1.97)
Overall Combined: Obama 42.58, Romney 44.72; Romney +2.14 (MOE: 1.32)




So what the hell does all that mean? Well you can interpret them as you wish but I would suggest looking first at Adjusted Average – Verified Professional as a baseline because that takes the 32/32/36 model out of the equation and it weights the results according to sample size which is very important. At this point that is Romney +2.40 (MOE: 1.77). Next I would look at Adjusted Average – Verified Combined and compare it to Adjusted Average – Converted Combined to gauge whether the conversion looks reasonable. There is a 0.64 difference…..pretty damned close. Then I would look at Adjusted Average – Overall Combined for the bottom line statistic which currently is Obama 42.58, Romney 44.72; Romney +2.14 (MOE: 1.32). What that tells us is that if the election was held last week Romney would probably win a squeaker in the popular vote as it is outside the margin of error.

Now as far as the trends are concerned, I went back and did this for each week since we started. Romney has held a similar lead for the last four weeks with a high spread of 2.65% (MOE: 1.38) and a low of 2.14% (MOE: 1.64) for the week ending 5/15. Both trendlines are descending, however at almost exactly the same rate meaning that undecided voters are increasing. Projected through election day according to the LV data we have now Romney would win the popular vote by roughly 3.50%

Why are you using 32 -32 D's and R's when 2008 was 39 to 32 D's over R's?

 

[NYcarbineer]

Obama needs to win what Kerry won,

plus Ohio,

and he's very likely to win Ohio.

 

[NYcarbineer]

Obama hit 51% approval on Rasmussen today, the first time since February. Based on Rasmussen's methodology,

that translates into about a 55% approval in the real world of polling.

 

[NYcarbineer]

It appears that our resident poll analysts have lost their enthusiasm for poll analyzing.

Is that because conservative BFF Rasmussen now has Obama ahead of Romney?

Is that because Raz has Obama at 50% approval? That's a virtually unbeatable number for a presidential election.

Obama's Approval Ratings Suggest a Nail-Biter - NYTimes.com

Historically, the best predictor of a president’s re-election chances has been approval rating. Since World War II, every president with an approval rating at least a few points above 50 percent has won re-election. Every president with a rating clearly below 50 percent has lost.

Mr. Obama now hovers in the gray area between likely victory and likely defeat, with the country divided nearly 50-50 on his performance. It is roughly the same place where Harry S. Truman, Gerald R. Ford and George W. Bush were when they ran as incumbents. Mr. Truman and Mr. Bush won close races, while Mr. Ford lost one.

Doesn't sound virtually unbeatable to me.

Ford wasn't up for re-election. He wasn't elected the first time.

 

[BluePhantom]

Well I think I have found a nominee for the "Blue Middle Finger of the Fucking Year" award. A project new america / keating (D) poll of Colorado showing Obama with a 4 point lead. Now what makes this such a fucking disgrace is they provide this fucking memo as a polling release. No demographic breakdowns, no information at all about how many Democrats were surveyed vs. Republicans, etc. No crosstabs to speak of, no definition of what an "unaffiliated voter" is. They imply it's Independents but most polls put unaffiliated and Independents in separate categories. And they have done only 20 polls in six years. Gee! Note the contact....Ethan Axelrod....yes that's David Axelrod's son.

Now who the hell is Keating Research? Beats me. I have never heard of them. I can't find anything on them earlier than late 2011. The founder is Chris Keating and I can't find any info on him....their website is two whole pages, one of which is his contact information and the other is apparently his self-written bio. He is however strongly affiliated with the Democratic party as his bio indicates his work in support of Democratic candidates.

Now my favorite part:

In the critical state of Colorado, Barrack Obama currently has a solid 4-point lead over Republican Mitt Romney.....

then below:

The margin of error for the survey is +/- 4.0%

In other words it's a statistical tie.

Now unfortunately I have to calculate this turd into the spreadsheet as is, but I'll tell you...this has got to be the poster child for "bullshit propaganda polls".

 

[saveliberty]

Better make a spreadsheet for votes saved/created...

 

[Foxfyre]

Well I think I have found a nominee for the "Blue Middle Finger of the Fucking Year" award. A project new america / keating (D) poll of Colorado showing Obama with a 4 point lead. Now what makes this such a fucking disgrace is they provide this fucking memo as a polling release. No demographic breakdowns, no information at all about how many Democrats were surveyed vs. Republicans, etc. No crosstabs to speak of, no definition of what an "unaffiliated voter" is. They imply it's Independents but most polls put unaffiliated and Independents in separate categories. And they have done only 20 polls in six years. Gee! Note the contact....Ethan Axelrod....yes that's David Axelrod's son.

Now who the hell is Keating Research? Beats me. I have never heard of them. I can't find anything on them earlier than late 2011. The founder is Chris Keating and I can't find any info on him....their website is two whole pages, one of which is his contact information and the other is apparently his self-written bio. He is however strongly affiliated with the Democratic party as his bio indicates his work in support of Democratic candidates.

Now my favorite part:

In the critical state of Colorado, Barrack Obama currently has a solid 4-point lead over Republican Mitt Romney.....

then below:

The margin of error for the survey is +/- 4.0%

In other words it's a statistical tie.

Now unfortunately I have to calculate this turd into the spreadsheet as is, but I'll tell you...this has got to be the poster child for "bullshit propaganda polls".

LOL, no kidding. Give me the money Axelrod probably paid Project New America to do this (cough) poll, and I'll give you authentic results showing that Kermit the Frog will be the next President of the United States. Most especially if he chooses Miss Piggy as his running mate.

And come on Phantom. This isn't rocket science.

"Unaffiliated voters" in this case is no doubt a group of college underclassmen attending a Gay Nazi rally or some such on campus. Or maybe all the folks waiting to see their 'counselor' at the local welfare office.

 

[BluePhantom]

Well I think I have found a nominee for the "Blue Middle Finger of the Fucking Year" award. A project new america / keating (D) poll of Colorado showing Obama with a 4 point lead. Now what makes this such a fucking disgrace is they provide this fucking memo as a polling release. No demographic breakdowns, no information at all about how many Democrats were surveyed vs. Republicans, etc. No crosstabs to speak of, no definition of what an "unaffiliated voter" is. They imply it's Independents but most polls put unaffiliated and Independents in separate categories. And they have done only 20 polls in six years. Gee! Note the contact....Ethan Axelrod....yes that's David Axelrod's son.

Now who the hell is Keating Research? Beats me. I have never heard of them. I can't find anything on them earlier than late 2011. The founder is Chris Keating and I can't find any info on him....their website is two whole pages, one of which is his contact information and the other is apparently his self-written bio. He is however strongly affiliated with the Democratic party as his bio indicates his work in support of Democratic candidates.

Now my favorite part:

In the critical state of Colorado, Barrack Obama currently has a solid 4-point lead over Republican Mitt Romney.....

then below:

The margin of error for the survey is +/- 4.0%

In other words it's a statistical tie.

Now unfortunately I have to calculate this turd into the spreadsheet as is, but I'll tell you...this has got to be the poster child for "bullshit propaganda polls".

LOL, no kidding. Give me the money Axelrod probably paid Project New America to do this (cough) poll, and I'll give you authentic results showing that Kermit the Frog will be the next President of the United States. Most especially if he chooses Miss Piggy as his running mate.

And come on Phantom. This isn't rocket science.

"Unaffiliated voters" in this case is no doubt a group of college underclassmen attending a Gay Nazi rally or some such on campus. Or maybe all the folks waiting to see their 'counselor' at the local welfare office.

oh yeah you're right on the damned nose and the real piss off is that since there has not been a poll of Colorado since we started on April 10th this piece of shit is going to be the baseline. It going to totally fuck up the trendline for at least a month. I mean we're not going to REALLY get a feel for Colorado until about late August now unless a plethora of Colorado polls comes rolling in over the next couple weeks. This kind of shit just annoys the living hell out of me....petaQ'pu' beQa!!!!!

 

[BluePhantom]

The weekly report will be delayed today due to the passing of a family member. I will try to get it done by tonight.

 

[Foxfyre]

So sorry for your loss Phantom. We'll be here when you are able to return.

 

[tjvh]

Obama hit 51% approval on Rasmussen today, the first time since February. Based on Rasmussen's methodology,

that translates into about a 55% approval in the real world of polling.

Interesting in that about 51% seem to be on Government handout programs. But I "sincerely" doubt there's a connection.

 

[saveliberty]

My taxes are up under Obama. I suppose people will take a close look at their situation just before the election and the election will swing accordingly.

 

[Foxfyre]

Some headlines from Rasmussen's site today:

38% View Federal Reserve Favorably, 49% Do Not
46% See Obama-Romney Race as Choice of Lesser of Two Evils
38% Rate Their Personal Finances As Good or Excellent
'Austerity' Talk Is Just Political Cover for More Government Spending By Scott Rasmussen
48% Agree With Romney on Issues, 46% Say Same of Obama

But in Rasmussen's Presidential assessment today: Obama's approval is somewhat 49% favorable and 49% unfavorable and Romney and Obama are each attracting 45% of the vote.

More ominous for Romney at this time is Rasmussen's assessment that Obama supporters are more excited and enthusiastic about their candidate than are Romney supporters. This indicates that Romney is going to have to work harder to reveal who he is to the American people and earn their trust and affection.

 

[The T]

Some headlines from Rasmussen's site today:

38% View Federal Reserve Favorably, 49% Do Not
46% See Obama-Romney Race as Choice of Lesser of Two Evils
38% Rate Their Personal Finances As Good or Excellent
'Austerity' Talk Is Just Political Cover for More Government Spending By Scott Rasmussen
48% Agree With Romney on Issues, 46% Say Same of Obama

But in Rasmussen's Presidential assessment today: Obama's approval is somewhat 49% favorable and 49% unfavorable and Romney and Obama are each attracting 45% of the vote.

More ominous for Romney at this time is Rasmussen's assessment that Obama supporters are more excited and enthusiastic about their candidate than are Romney supporters. This indicates that Romney is going to have to work harder to reveal who he is to the American people and earn their trust and affection.

And Certain Americans to not belive everything they hear from the media and look for themselves. It's early yet.

 

[The T]

The weekly report will be delayed today due to the passing of a family member. I will try to get it done by tonight.

Take your time friend. Take care of business first!

Sorry for your loss.

 

[BluePhantom]

Weekly Breakdown for 5/29/12

It seems I might be able to bang this out relatively quickly after all since there was frankly a whole lot of nothing last week to look at. Only Gallup and Rasmussen’s tracking polls were released last week for the general election although there were some state polls to look at, but unfortunately, there wasn’t much that was reliable. What this essentially means is we will have no LV Report this week nor a Demographics Report since there’s just not much data to crunch.
With that said let's look at the statistics we DO have:

5/22/12 - 5/29/12

Base Average
Professional Polls
Rasmussen (LV - tracking): Obama 45, Romney 45
Gallup (RV - tracking): Obama 47, Romney 45

Professional Average: Obama 46.00, Romney 45.00; Obama +1.00

Media Polls
None

Media Average: N/A

Combined Base Average: Obama 46.00, Romney 45.00; Obama +1.00 (MOE: 1.49)


Adjusted Average
Professional Polls
Rasmussen (LV - tracking): Obama 675, Romney 675
Gallup (RV - tracking): Obama 1034, Romney 990

Total: Obama 1709, Romney 1665

Professional Average: Obama 46.19, Romney 45.00; Obama +1.19 (MOE: 1.47)

Media Polls
None

Total: N/A

Media Average: N/A

Combined Adjusted Average: Obama 46.19, Romney 45.00; Obama +1.19 (MOE: 1.47)


Two Week Rolling Average

Base Average
Professional Polls

Obama 45.17, Romney 44.50; Obama +0.67

Media Polls

Obama 47.33, Romney 42.67; Obama +4.67

Combined Average

Obama 45.89, Romney 43.89; Obama +2.00

Adjusted Average
Professional Polls

Obama 45.89, Romney 45.41; Obama +0.49

Media Polls

Obama 45.73, Romney 43.72; Obama +2.01

Combined Average

Obama 45.85, Romney 45.02; Obama +0.84

In state polls there was some more information. A Project New America/Keating (D) LV poll of Colorado projects Obama with a 48/44 lead. I have already made reference to this poll in a different post but since that’s all we have, that’s what we must go with. It’s important to note that the MOE was 4 so Colorado is a statistical tie.

NBC News/Marist released three RV state polls this week all showing Obama leads beyond the trend and the averages of other pollsters. In Florida they calculate a 48/44 Obama lead which is totally contrary to the trend and Quinnipiac’s survey of Florida last week. In Ohio they calculate a 48/42 Obama lead which is again above the trend and average and in Virginia they report an Obama lead of four points which again would be above the trend and averages. Unfortunately they did not provide crosstabs on any of these polls so I can’t convert them into an LV model.

In North Carolina the Civitas (R) LV poll showed a 45/47 Romney lead. This is pretty much in line with previous data from SUSA last week and would fall in line with the trend.

In Pennsylvania Rasmussen’s LV poll showed a 47/41 Obama advantage which again is pretty much in line with everything else we have seen from the state once you adjust for RV vs. LV polls.

Finally in Wisconsin, a Wisconsin Public Radio LV poll shows Obama with a 49/43 lead. While this will be calculated into the spreadsheet I urge a LOT of caution with this poll. The sample size was 406, the MOE was a whopping 5 points, and they provided no demographics or crosstabs at all. A six point Obama lead would be WAY outside the trend and the averages…I mean not even in the same ballpark.

So I am sorry I don’t have time to get into more detail this week, but as I mentioned previously there’s just too much going on right now and fortunately there wasn’t crap as far as data anyhow. Hopefully, next week will be better in both regards.

 

[BluePhantom]

More ominous for Romney at this time is Rasmussen's assessment that Obama supporters are more excited and enthusiastic about their candidate than are Romney supporters. This indicates that Romney is going to have to work harder to reveal who he is to the American people and earn their trust and affection.

I would tend to agree with that assessment based upon what I have seen from the LV crosstabs. Currently Democratic voters seem more energized. I am not how sure that will last though. I just don't have enough data to establish a trendline in regard to voter enthusiasm that I would be very confident about. That will change in time though. The GOP is not yet united but once the debates start and Romney picks a running mate and all that they will come together and Republican enthusiasm will increase.

 

[NYcarbineer]

Weekly Breakdown for 5/29/12

It seems I might be able to bang this out relatively quickly after all since there was frankly a whole lot of nothing last week to look at. Only Gallup and Rasmussen’s tracking polls were released last week for the general election although there were some state polls to look at, but unfortunately, there wasn’t much that was reliable. What this essentially means is we will have no LV Report this week nor a Demographics Report since there’s just not much data to crunch.
With that said let's look at the statistics we DO have:

5/22/12 - 5/29/12

Base Average
Professional Polls
Rasmussen (LV - tracking): Obama 45, Romney 45
Gallup (RV - tracking): Obama 47, Romney 45

Professional Average: Obama 46.00, Romney 45.00; Obama +1.00

Media Polls
None

Media Average: N/A

Combined Base Average: Obama 46.00, Romney 45.00; Obama +1.00 (MOE: 1.49)


Adjusted Average
Professional Polls
Rasmussen (LV - tracking): Obama 675, Romney 675
Gallup (RV - tracking): Obama 1034, Romney 990

Total: Obama 1709, Romney 1665

Professional Average: Obama 46.19, Romney 45.00; Obama +1.19 (MOE: 1.47)

Media Polls
None

Total: N/A

Media Average: N/A

Combined Adjusted Average: Obama 46.19, Romney 45.00; Obama +1.19 (MOE: 1.47)


Two Week Rolling Average

Base Average
Professional Polls

Obama 45.17, Romney 44.50; Obama +0.67

Media Polls

Obama 47.33, Romney 42.67; Obama +4.67

Combined Average

Obama 45.89, Romney 43.89; Obama +2.00

Adjusted Average
Professional Polls

Obama 45.89, Romney 45.41; Obama +0.49

Media Polls

Obama 45.73, Romney 43.72; Obama +2.01

Combined Average

Obama 45.85, Romney 45.02; Obama +0.84

In state polls there was some more information. A Project New America/Keating (D) LV poll of Colorado projects Obama with a 48/44 lead. I have already made reference to this poll in a different post but since that’s all we have, that’s what we must go with. It’s important to note that the MOE was 4 so Colorado is a statistical tie.

NBC News/Marist released three RV state polls this week all showing Obama leads beyond the trend and the averages of other pollsters. In Florida they calculate a 48/44 Obama lead which is totally contrary to the trend and Quinnipiac’s survey of Florida last week. In Ohio they calculate a 48/42 Obama lead which is again above the trend and average and in Virginia they report an Obama lead of four points which again would be above the trend and averages. Unfortunately they did not provide crosstabs on any of these polls so I can’t convert them into an LV model.

In North Carolina the Civitas (R) LV poll showed a 45/47 Romney lead. This is pretty much in line with previous data from SUSA last week and would fall in line with the trend.

In Pennsylvania Rasmussen’s LV poll showed a 47/41 Obama advantage which again is pretty much in line with everything else we have seen from the state once you adjust for RV vs. LV polls.

Finally in Wisconsin, a Wisconsin Public Radio LV poll shows Obama with a 49/43 lead. While this will be calculated into the spreadsheet I urge a LOT of caution with this poll. The sample size was 406, the MOE was a whopping 5 points, and they provided no demographics or crosstabs at all. A six point Obama lead would be WAY outside the trend and the averages…I mean not even in the same ballpark.

So I am sorry I don’t have time to get into more detail this week, but as I mentioned previously there’s just too much going on right now and fortunately there wasn’t crap as far as data anyhow. Hopefully, next week will be better in both regards.

Why did you leave out the NBC Colorado poll?