The Phantom's Polling Analysis Thread

[BluePhantom]

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.

Yes. I noticed that on the 7th Romney and Obama were tied in Rasmussen's tracking poll. On the 8th Romney had a five point lead and that held today as well. There was a similar jump in Gallup's tracking poll. Now Rasmussen uses a three day rolling average and Gallup a five day so looking at Rasmussen what that means is that their Tuesday results included the average of polls from Saturday, Sunday, and Monday. That means one of two things: either a) a massive Obama poll was taken on Friday and it fell off the rolling average or b) there was a massive Romney poll on Monday. We won't know for certain until Friday when the Monday poll falls off the average and depending on which way they go from here can give some indication whether it's noise or a real bump.

I speculate it's a trend because the other polls seem to be reinforcing that idea but also remember that the jobs report was released on Friday and it was dismal. It would make a lot of sense for the polls over the following three days to show a significant drop in Obama support as a result of that jobs report.

 

[BluePhantom]

Agency Average Deviation Report 5/10/2012


So per my previous post I did some number crunching and came up with some interesting data on house biases for the pollsters. Now, I have to say up front that this was not the most in depth statistical analysis in the world for a couple reasons: a) the data is spread all over the place and in some situations there was not enough other data to form a reasonable basis for comparison. In those cases I ignored the poll because the purpose is to determine how far each pollster deviates from the averages. So for example in New Mexico there has only been one poll since 4/10/12. Well...as there is no basis for comparison, there's nothing to do really but ignore that one. b) sometimes there were gaps in polling. Florida, for example, seems to get polled every other week by everyone and totally ignored the next week. So it became difficult to create a consistent statistical formula that would cover all the bases. Here's what I decided to do.

For each pollster where a current basis for comparison could be established (i.e. enough polls that particular week to compare the deviation directly according to the same time frame) I simply took their point spread and compared it to the average. When there were not enough polls in a given week to make that direct comparison I compared each poll to the trendline. When there was not enough data to establish a trendline, I ignored the poll. Is that incredibly sound statistical analysis? Well...no not really ....but it gives us a basic idea and that's all we're really after.

So here's what we have in descending order of proximity to the average. All pollsters use RV method unless otherwise noted. I will provide the pollster's name, the number of polls that qualified according to the above criteria, and their current "lean" based on the above method. I must admit I saw some surprising rankings. Keep in mind this is NOT an evaluation of "accuracy". That can't yet be determined and there are some things to really keep in mind such as pollsters that only had one qualifying poll and most certainly the RV/LV effect which will push the average several points to the left and so when you compare an LV poll to a large group of RV polls one would expect a higher deviation from the average. All this measures (and a very basic measurement at that) is "to this point, how far away have their results been compared to the averages of all results?"

1) Fox News: 3 polls - Romney +0.33
2) Gallup: 6 polls - Obama +0.36
3) Purple Strategies: 1 poll - Romney +1
4) Politico: 1 poll - Obama +1.67
5) Rasmussen (LV): 7 polls - Romney +2.26
6) Democracy Corps (D) (LV*): 1 poll - Obama +2.67
7) Quinnipiac: 1 poll - Obama +4.00
8) ABC/Washington Post: 1 poll - Obama +4.00
9) PPP (D): 4 polls - Obama +5.00
10) TIPPOnline: 1 poll - Obama +5.67
11) National Journal: 1 poll - Obama +8.00
12) Reuters/Ipsos: 1 poll - Obama +9.67
13) Associated Press: 1 poll - Obama +12.00

*Listed officially as LV but is RV in reality

Now you are free to interpret this as you wish. I personally would ignore all the listings where there was only one poll because sometimes you get lucky and sometimes you get an outlier so even with the last three polls that are way the fuck out there, I personally will give them the benefit of the doubt and say "maybe you had a bad day". Additionally the top six are all within the MOE so no worries there. If you choose to do that, then you are left with this:

1) Fox News: 3 polls - Romney +0.33
2) Gallup: 6 polls - Obama +0.36
3) Rasmussen (LV): 7 polls - Romney +2.26
4) PPP (D): 4 polls - Obama +5.00

Now the Rasmussen 2.26 deviation again makes sense since they are LV polls compared to a massive sample of RV polls which is going to push the average (as an overall average it will probably push it 3-4 points in Obama's favor...for an individual poll somewhere between 4-6 points but as an average that effect will be mellowed a little bit) so if you apply that effect and compensate, Rasmussen is right about where they should be. But as I suspected...what the hell is up with PPP? An average of four polls outside the MOE in three total weeks where they had a qualifying poll?!?!?!? Yikes! Add in that of those four pollsters, PPP is the only one to show an Obama lean in every single poll they released. There rest had weeks when their data leaned toward Romney against the average and weeks when their data leaned toward Obama against the average or at least hit the average on the nose.

Now as I said earlier...perhaps they are right and everyone else is wrong and I am not prepared to accuse them of intentional wrongdoing, but there is something happening in their polling method that is consistently producing results that lean toward Obama. For now I will give them the benefit of the doubt that this is an unintended effect because historically they have given me little to no reason to suggest otherwise. But I think the results I have shown warrant an increased degree of skepticism where their data is concerned.

 

[Foxfyre]

Which looks reasonable Phantom. Today's updated RCP average based on the most recent polls they are tracking has Obama up 1.3%.

Again it is too early to draw any firm conclusions, but I am thinking the numbers do not look good for Obama, most especially if the economy does not significantly improve. Obama's opponent in the West Virginia primary was a guy who is in a Texas prison. And he got 40% of the Democratic vote. When you see numbers like that, you wonder how strong the lean toward Barack Obama really is.

 

[BluePhantom]

Which looks reasonable Phantom. Today's updated RCP average based on the most recent polls they are tracking has Obama up 1.3%.

Which is reasonable. RCP calculates it a bit differently than we are doing here. They only do what we are calling "Overall Base Average (OBA)" and they keep their polls into the average for a longer time than we do so there will be some slight differences. As of this minute my spreadsheet shows Obama by 0.25% according to OBA...so yeah...their 1.3% is a reasonable OBA result. When you start adjusting for sample size, making separate calculations of media polls and professional polls, and all the stuff we do here it starts to lean more toward Romney, but not too terribly much...our current Adjusted Professional Average (APA) has Romney by 1.33% so even the spread between RCP's OBA and our APA is within the margin of error.

In other words we pretty much all agree

 

[BarbaraHavers]

Which looks reasonable Phantom. Today's updated RCP average based on the most recent polls they are tracking has Obama up 1.3%.

Again it is too early to draw any firm conclusions, but I am thinking the numbers do not look good for Obama, most especially if the economy does not significantly improve. Obama's opponent in the West Virginia primary was a guy who is in a Texas prison. And he got 40% of the Democratic vote. When you see numbers like that, you wonder how strong the lean toward Barack Obama really is.

Who's crazier, Keith Judd or the people who voted for him? In any event, Obama still got almost 30K votes more than Romney, who was the Republican front runner.

Here's a link to Keith Judd and some of his positions. He is

"As for his priority issues, Judd lists the Judicial Accountability Commission and U.S. Department of Justice Accountability Commission as well as the U.S. Treasury Accountability Commission, the Federal Reserve Bank System Accountability Commission. He'd also like to "re-invent the public school system with ties to industries, community job placement and higher education."

He seeks office "to make the world a better place."

Other interesting facts from Judd's profile:

* He lists his mother as actress Lillian Russell, a silent film star who died in 1922;

* He also says his father, Homer. T. Judd, "designed the first atomic bomb and worked for the Atomic Energy Commission 30 years;"

* His religion: "Rastafarian-Christian;"

* He also says he was a member of the Federation of Super Heroes from 1976 to 1982;

* Under education, he said he's "attended" The Kennedy School of Government at Harvard, UCLA, the University of New Mexico and Bethany Nazarene College.

* He wants to eliminate federal taxes and increase spending on defense and domestic welfare. "The Federal Government can merely operate on its own self-produced money rather than pretend to borrow it to pay private banks to store it and regulate interest rates, and then tax it;"

* He doesn't believe in homosexuality. "Homosexuality does not exist anywhere else in nature. No other creature or animal on earth are homosexuals. I do not condone crimes or attacks on homosexuals, but we do not have to place our seal of approval of same sex marriages."

Inmate gets 41 percent of W.Va. vote against Obama | West Virginia Record

 

[BluePhantom]

Which looks reasonable Phantom. Today's updated RCP average based on the most recent polls they are tracking has Obama up 1.3%.

Again it is too early to draw any firm conclusions, but I am thinking the numbers do not look good for Obama, most especially if the economy does not significantly improve. Obama's opponent in the West Virginia primary was a guy who is in a Texas prison. And he got 40% of the Democratic vote. When you see numbers like that, you wonder how strong the lean toward Barack Obama really is.

Who's crazier, Keith Judd or the people who voted for him? In any event, Obama still got almost 30K votes more than Romney, who was the Republican front runner.

Here's a link to Keith Judd and some of his positions. He is

"As for his priority issues, Judd lists the Judicial Accountability Commission and U.S. Department of Justice Accountability Commission as well as the U.S. Treasury Accountability Commission, the Federal Reserve Bank System Accountability Commission. He'd also like to "re-invent the public school system with ties to industries, community job placement and higher education."

He seeks office "to make the world a better place."

Other interesting facts from Judd's profile:

* He lists his mother as actress Lillian Russell, a silent film star who died in 1922;

* He also says his father, Homer. T. Judd, "designed the first atomic bomb and worked for the Atomic Energy Commission 30 years;"

* His religion: "Rastafarian-Christian;"

* He also says he was a member of the Federation of Super Heroes from 1976 to 1982;

* Under education, he said he's "attended" The Kennedy School of Government at Harvard, UCLA, the University of New Mexico and Bethany Nazarene College.

* He wants to eliminate federal taxes and increase spending on defense and domestic welfare. "The Federal Government can merely operate on its own self-produced money rather than pretend to borrow it to pay private banks to store it and regulate interest rates, and then tax it;"

* He doesn't believe in homosexuality. "Homosexuality does not exist anywhere else in nature. No other creature or animal on earth are homosexuals. I do not condone crimes or attacks on homosexuals, but we do not have to place our seal of approval of same sex marriages."

Inmate gets 41 percent of W.Va. vote against Obama | West Virginia Record

Wow....what a fruitcake. And he won 10 of 55 counties. Unreal.

Yeah I agree with Firefox that the numbers are not promising for Obama. Certainly the current trends suggest that. The projected trendline on my spreadsheet is above what I call "the hard deck" and Obama's is below "the hard deck". What that means (it's hard to explain in detail without seeing the chart) is that at current pace and trends, Romney's support is rising and is expected to gain momentum, while Obama's support is falling and is expected to lose momentum.

Now I am not going to profess that with a great deal of tenacity right now because we have only been calculating for a month, it's still so far away from election day that the trends can (and will) change in some way, and so many things can happen (especially with the economy), that I am loathe to get too carried away with it all. What I would say is that if someone put a gun to my head right now and told me to predict what will happen in November, I would bet on a Romney win based on the data I currently have and the trends they suggest.

 

[Foxfyre]

I won't comment on it further lest we derail Phantom's fine thread here, Barbara, but the man is in prison convicted for a murder he committed here in New Mexico. The only reason 41% of WV voters voted for him was that was the only way they could repudiate Obama. Judd could have been running on an ice cream and bunny rabbits platform and he would still have pulled votes.

The other candidates are still pulling votes in the Republican Primary of course which affects the percentages. But you add that 41% who are saying they probably won't vote Obama to the votes for Romney plus he'll also get most of the votes that went to other candidates, and Obama doesn't look too good in West Virginia does he. Most especially when last month's polls had Romney at 54% to Obama 37% in a head to head matchup.

 

[BluePhantom]

Normally I would wait until next week to refer to this but there is some very interesting information out today. First, as we discussed earlier we would need to wait until Friday (today) to determine why the Rasmussen tracking poll showed a significant jump from Monday to Tuesday's results. We can now conclude that if there was an outlier poll in favor of Obama it would have occurred last Friday and because the Rasmussen polls remained relatively stable until today that there was not a massive Romney poll skewing the numbers. Today, Rasmussen's results have shown an even greater increase in Romney support (a 7 point spread). This indicates that according strictly to Rasmussen's numbers the jump in Romney support from last week is the continuation of a trend that is picking up momentum. Gallup's results so far neither support or refute that idea (remember they run a five day average compared to Rasmussen's three) and no other polls from professional pollsters have been released in the last couple days in regard to the national numbers. However as things stand today, the average of professional pollsters shows the first occasion where Romney is now holding a lead which is outside the margin of error.

What is very alarming is a SUSA RV poll released today showing Obama with a mere four point lead in Oregon. SUSA is the only organization to poll Oregon this year (once in March and once this month). According to their numbers Obama's support in Oregon is dropping like a stone. Now we have to be careful here. There are no other polls to support or refute SUSA's data so that demands caution....even SUSA is not immune to the occasional outlier. However, I have SUSA ranked #1 on my list of pollster accuracy (and most other rankings agree that they are at least in the top three) so their data carries weight. If this poll is correct....less me stress the word "if"....then Oregon is in play and that represents an ominous sign for Obama supporters as Oregon is one damned liberal state....trust me...I live here. Also as this was an RV poll which on average will give a free 4-6 points to the Democratic candidate it implies that Romney is in a dead heat and maybe even holds a slight lead.

Now before I get all hot and bothered and put Oregon in the "swing state" category I need to see other polls that support this data. This is just one poll so for now I am leaving Oregon in Obama's EV count, much like I need to see more polls of New Mexico before I remove it from "swing state" status and move it into Obama's column. But SUSA's history of accuracy makes this poll worth mention and something to keep an eye on.

 

[Foxfyre]

Velly interesting Phantom. I wonder how much the dynamics within the individual states are forecasters of what would really happen in the general election if that were held today. (We have agreed that today's polling may or may not have any bearing on how the vote will go in November.)

But the dynamics are fascinating. For instance in Wisconsin, the most interesting state to me right now, Rasmussen has Walker 50% - Barnett 45% in the recall election matchup and in the Senate the Republican leads the Democrat 50% to 35%. But Obama is at 49% and Romeny 45%. Doesn't really compute does it? Or does it?

 

[BluePhantom]

Velly interesting Phantom. I wonder how much the dynamics within the individual states are forecasters of what would really happen in the general election if that were held today. (We have agreed that today's polling may or may not have any bearing on how the vote will go in November.)

But the dynamics are fascinating. For instance in Wisconsin, the most interesting state to me right now, Rasmussen has Walker 50% - Barnett 45% in the recall election matchup and in the Senate the Republican leads the Democrat 50% to 35%. But Obama is at 49% and Romeny 45%. Doesn't really compute does it? Or does it?

If you look at the trends it actually does. The most recent poll in Wisconsin, aside from today's Rasmussen data, was a Marquette University poll showing Obama with a nine point lead. Now Marquette is a little odd. They are not media but they are not professional either so there's a bit of a gray area on exactly what kind of pollster they are. But if you compare the two (and accept both as accurate) then it suggests a five point swing in Romney's favor in Wisconsin over the last two weeks. Thompson and Walker have both seen similar surges in Wisconsin over the same time frame. So those other polls would certainly support the idea of Romney closing the gap in Wisconsin

 

[Foxfyre]

Sorry about the post I just deleted if anybody saw it. It was intended for the Coffee Shop.

 

[The T]

Sorry about the post I just deleted if anybody saw it. It was intended for the Coffee Shop.

No harm, no foul.

 

[BluePhantom]

Weekly Breakdown for 5/15/12

An interesting set of polling data this week. For the most part everyone seemed to agree except the AP poll that was significantly off from everyone else. A quick look at their data and it’s pretty easy to speculate as to why but I will cover that in another post. Unfortunately, we have about the weakest set of crosstabs and demographics data to work with so the demographics and trends report will again be pretty iffy (assuming I can squeeze any data out of them at all).

On an interesting note RCP chose not to calculate the CBS/NYT poll in their overall overage as it is a call-back poll. Basically they are comparing previous results with the same people they contacted before to see if opinions have changed among the same sample. That’s an interesting comparison we will look closely at in another post. As it is still a recent poll and the previous poll has fallen out of our calculation range we will include the poll in our data set. Rasmussen and Gallup’s tracking polls showed some dramatic movement. Both showed heavy movement toward Romney over the weekend but in the last two days both their gaps have closed. At one point Romney held a 7 point spread on Rasmussen’s poll but now holds only a two point advantage. As usual all links can be found on RCP.
With that said let's look at the statistics:

5/9/12 - 5/15/12

Base Average
Professional Polls
Rasmussen (LV - tracking): Obama 45, Romney 47
Gallup (RV - tracking): Obama 46, Romney 45

Professional Average: Obama 45.50, Romney 46.00; Romney +0.50

Media Polls
CBS/NYT (RV): Obama 43, Romney 46
Associated Press (RV): Obama 50, Romney 42
Washington Times (LV): Obama 43, Romney 44

Media Average: Obama 45.33, Romney 44.00; Obama +1.33

Overall Base Average: Obama 45.40, Romney 44.80; Obama +0.60


Adjusted Average
Professional Polls
Rasmussen (LV - tracking): Obama 675, Romney 705
Gallup (RV - tracking): Obama 1012, Romney 990

Total: Obama 1687, Romney 1695

Professional Average: Obama 45.59, Romney 45.81; Romney +0.22

Media Polls
CBS/NYT (RV): Obama 241.66, Romney 258.52
Associated Press (RV): Obama 435.50, Romney 365.82
Washington Times (LV): Obama 344, Romney 352

Total: Obama 1021.20, Romney 976.34

Media Average: Obama 45.73, Romney 43.72; Obama +2.01
Overall Adjusted Average: Obama 45.65, Romney 45.03; Obama +0.62


Two Week Rolling Average

Base Average
Professional Polls

Obama 45.20, Romney 47.00; Romney +1.80

Media Polls

Obama 46.33, Romney 44.17; Obama +2.17

Combined Average

Obama 45.82, Romney 45.45; Obama +0.36

Adjusted Average
Professional Polls

Obama 45.06, Romney 46.83; Romney +1.77

Media Polls

Obama 46.65, Romney 44.12; Obama +2.53

Combined Average

Obama 45.66, Romney 45.81; Romney +0.16

In state polls there was some more bad news for Romney continuing over from last week. A Suffolk/Channel 7 News LV poll of Florida gave a 46/45 advantage to Obama. If accurate this would represent a two point swing to Obama since the previous time the state was polled two weeks ago. I am not sure how much I trust “Channel 7 News” to develop accurate polling methods but we will give them the benefit of the doubt for now.

In New Hampshire a PPP (D) RV poll showed Obama with a 53/41 lead over Romney. As we discussed last week PPP is on the “view with skepticism” list right now so take it as you wish. Unfortunately there are no other polls that were taken after April 10th to compare it to so who really knows. Time will tell on that one.

In Ohio the Quinnipiac RV poll showed a 45/44 Obama advantage. This supports earlier polls by Rasmussen and Quinnipiac that showed very similar spreads and casts further doubt on PPP (D)’s poll last week showing a 7 point Obama advantage.

In Wisconsin we have two polls this week. Rasmussen’s LV poll shows a four point Obama lead 49/45, but the Daily Kos/PPP (D) poll has it at a 47/46 Obama advantage. This is very confusing as it’s the first time in weeks that PPP has released a poll that was not completely out of synch with the majority of other pollsters and produced a poll where Romney actually performed better. The affiliation with the Daily Kos in this particular poll might have something to do with that, but that’s speculation. Regardless of how you choose to look at them it creates a base average of Obama +2.50 and an adjusted average spread of Obama +2.11. All polls are within the MOE or right on the edge of it. This is worth keeping a very close eye on as technically we have a statistical tie in Wisconsin with a trend toward Romney. Wisconsin is definitely in play right now.

 

[BluePhantom]

Just a note on this Daily Kos/PPP(D) poll of Wisconsin. This is very very very very interesting. DK fired Research2000 as their polling agency in 2010 after a disastrous showing of inaccuracy by R2K. DK's political biases are quite apparent and as they released poll after poll showing heavy Democratic leads (which turned out to be completely wrong) they were facing heavy criticism of intentional wrongdoing. So out went R2K and in came PPP. What is so interesting about this is that regardless of any opinion you may (or may not) have about DK's website biases, it stands to reason that they want accurate polling information else they would not have fired R2K.

With that rationale it seems amazing to me that the polls PPP has released that were funded by the DNC have been showing massive Obama leads, but suddenly they release a poll of Wisconsin commissioned by DK and it's right there with other pollsters. Well Sacrebleu!! Now before I go off half cocked and start making accusations I want to see if this continues to happen. If we see a pattern where the PPP polls commissioned by the DNC are showing huge Obama leads and the PPP polls commissioned by other organizations are in line with everyone else, there can only be one conclusion and I don't think I have to spell it out for you.

The Wisconsin poll results certainly suggest something like that may be happening, but again I want to see the same thing happen over several polls before I start making strong accusations of intentional tampering. For NOW all I will say is "hmmm....this looks fishy", but I am going to be watching this situation like a fucking hawk. I will report the results as they come.

 

[PoliticalChic]

Weekly Breakdown for 5/15/12

An interesting set of polling data this week. For the most part everyone seemed to agree except the AP poll that was significantly off from everyone else. A quick look at their data and it’s pretty easy to speculate as to why but I will cover that in another post. Unfortunately, we have about the weakest set of crosstabs and demographics data to work with so the demographics and trends report will again be pretty iffy (assuming I can squeeze any data out of them at all).

On an interesting note RCP chose not to calculate the CBS/NYT poll in their overall overage as it is a call-back poll. Basically they are comparing previous results with the same people they contacted before to see if opinions have changed among the same sample. That’s an interesting comparison we will look closely at in another post. As it is still a recent poll and the previous poll has fallen out of our calculation range we will include the poll in our data set. Rasmussen and Gallup’s tracking polls showed some dramatic movement. Both showed heavy movement toward Romney over the weekend but in the last two days both their gaps have closed. At one point Romney held a 7 point spread on Rasmussen’s poll but now holds only a two point advantage. As usual all links can be found on RCP.
With that said let's look at the statistics:

5/9/12 - 5/15/12

Base Average
Professional Polls
Rasmussen (LV - tracking): Obama 45, Romney 47
Gallup (RV - tracking): Obama 46, Romney 45

Professional Average: Obama 45.50, Romney 46.00; Romney +0.50

Media Polls
CBS/NYT (RV): Obama 43, Romney 46
Associated Press (RV): Obama 50, Romney 42
Washington Times (LV): Obama 43, Romney 44

Media Average: Obama 45.33, Romney 44.00; Obama +1.33

Overall Base Average: Obama 45.40, Romney 44.80; Obama +0.60


Adjusted Average
Professional Polls
Rasmussen (LV - tracking): Obama 675, Romney 705
Gallup (RV - tracking): Obama 1012, Romney 990

Total: Obama 1687, Romney 1695

Professional Average: Obama 45.59, Romney 45.81; Romney +0.22

Media Polls
CBS/NYT (RV): Obama 241.66, Romney 258.52
Associated Press (RV): Obama 435.50, Romney 365.82
Washington Times (LV): Obama 344, Romney 352

Total: Obama 1021.20, Romney 976.34

Media Average: Obama 45.73, Romney 43.72; Obama +2.01
Overall Adjusted Average: Obama 45.65, Romney 45.03; Obama +0.62


Two Week Rolling Average

Base Average
Professional Polls

Obama 45.20, Romney 47.00; Romney +1.80

Media Polls

Obama 46.33, Romney 44.17; Obama +2.17

Combined Average

Obama 45.82, Romney 45.45; Obama +0.36

Adjusted Average
Professional Polls

Obama 45.06, Romney 46.83; Romney +1.77

Media Polls

Obama 46.65, Romney 44.12; Obama +2.53

Combined Average

Obama 45.66, Romney 45.81; Romney +0.16

In state polls there was some more bad news for Romney continuing over from last week. A Suffolk/Channel 7 News LV poll of Florida gave a 46/45 advantage to Obama. If accurate this would represent a two point swing to Obama since the previous time the state was polled two weeks ago. I am not sure how much I trust “Channel 7 News” to develop accurate polling methods but we will give them the benefit of the doubt for now.

In New Hampshire a PPP (D) RV poll showed Obama with a 53/41 lead over Romney. As we discussed last week PPP is on the “view with skepticism” list right now so take it as you wish. Unfortunately there are no other polls that were taken after April 10th to compare it to so who really knows. Time will tell on that one.

In Ohio the Quinnipiac RV poll showed a 45/44 Obama advantage. This supports earlier polls by Rasmussen and Quinnipiac that showed very similar spreads and casts further doubt on PPP (D)’s poll last week showing a 7 point Obama advantage.

In Wisconsin we have two polls this week. Rasmussen’s LV poll shows a four point Obama lead 49/45, but the Daily Kos/PPP (D) poll has it at a 47/46 Obama advantage. This is very confusing as it’s the first time in weeks that PPP has released a poll that was not completely out of synch with the majority of other pollsters and produced a poll where Romney actually performed better. The affiliation with the Daily Kos in this particular poll might have something to do with that, but that’s speculation. Regardless of how you choose to look at them it creates a base average of Obama +2.50 and an adjusted average spread of Obama +2.11. All polls are within the MOE or right on the edge of it. This is worth keeping a very close eye on as technically we have a statistical tie in Wisconsin with a trend toward Romney. Wisconsin is definitely in play right now.

A note on "CBS/NYT (RV): Obama 43, Romney 46"

More than passing interesting that 'Morning Joe' MSNBC noted that even this poll was slanted toward Obama, using 36% Democrats, and 30% Republicans...


Your efforts notwithstanding, and appreciated, it seems anticlimactic..

...it's all over but the shouting....

...and leaping from tall buildings by the Left.

 

[BluePhantom]

This week's "Blue Middle Finger Award" from the Phantom go to two polls that really piss me off. They are the AP RV poll (Obama +8) and the Washington Times LV poll (Romney +1). Why you ask? Well the information for the AP poll doesn't ask who the respondent supports until question #38 and they have refused to release the first 37 questions. WTF?!?!?!? If you are going to release a poll that is so out of line with every else's results you had better damn well back up your data and hiding the first 37 questions that you asked aint exactly the way to inspire confidence that your results are unbiased. Secondly, while they provide good demographics information of the sample as a whole they render them completely useless because they provide no crosstabs. Ok so we know that the survey sample was 13% Hispanic. Great....the average is 8% so we know it's slightly oversampled for Hispanics, but how did those Hispanics respond? They don't say. So fuck you AP!

Now for the Washington Times they are even worse. They best they offered was a chart and a newspaper article. No link to the questions they asked, when they asked them, whether they were rotated, no detailed demographics breakdowns, no crosstabs. Hey Washington Times....fuck you too!!!

This weeks "Blue Middle Finger Awards" are brought to you by the Blue Phantom. They do not necessarily represent the views of USMB, their owners, or affiliates.

 

[BluePhantom]

A note on "CBS/NYT (RV): Obama 43, Romney 46"

More than passing interesting that 'Morning Joe' MSNBC noted that even this poll was slanted toward Obama, using 36% Democrats, and 30% Republicans...

Yes, but that's pretty standard with an RV poll. RV polls will nearly always have the Democrats oversampled because there are more registered Democrats in the United States but they vote are a far lower rate. So when the pollsters filtering question is simply "are you registered to vote" you are going to naturally reach more registered Democrats than Republicans. But because Republicans vote at a higher rate than Democrats, when it gets down to who actually shows up to vote it's usually about a 32%/32% split with Independents (who are the most active voters of all) making up the remaining 36%. So their sample of 36/30/34 (D/R/I) is about what you would expect from an RV poll.

Now just for shits and giggles let's take that and convert it into a generic 32/32/36 LV model and see what we get.

Sample size: 562

32% Democrats = 179.84
32% Republican = 179.84
36% Independent = 202.32

Democrats: Obama 84%, Romney 10% = Obama 151.07, Romney 17.984
Republicans: Obama 5%, Romney 91% = Obama 8.992, Romney 163.65
Independents: Obama 36%, Romney 43% = Obama 72.835, Romney 86.998

Total: Obama 232.89, Romney 268.64 = Obama 41.44%, Romney 47.80%; Romney +6.36

Your efforts notwithstanding, and appreciated, it seems anticlimactic..

...it's all over but the shouting....

...and leaping from tall buildings by the Left.

It's way too early to start declaring victory. It's certainly starting to trend strong toward Romney (and Obama's statement on gay marriage just handed Romney North Carolina on a silver platter) but Romney has some problems in a few key states. A lot can happen.

 

[PoliticalChic]

A note on "CBS/NYT (RV): Obama 43, Romney 46"

More than passing interesting that 'Morning Joe' MSNBC noted that even this poll was slanted toward Obama, using 36% Democrats, and 30% Republicans...

Yes, but that's pretty standard with an RV poll. RV polls will nearly always have the Democrats oversampled because there are more registered Democrats in the United States but they vote are a far lower rate. So when the pollsters filtering question is simply "are you registered to vote" you are going to naturally reach more registered Democrats than Republicans. But because Republicans vote at a higher rate than Democrats, when it gets down to who actually shows up to vote it's usually about a 32%/32% split with Independents (who are the most active voters of all) making up the remaining 36%. So their sample of 36/30/34 (D/R/I) is about what you would expect from an RV poll.

Now just for shits and giggles let's take that and convert it into a generic 32/32/34 LV model and see what we get.

Sample size: 562

32% Democrats = 179.84
32% Republican = 179.84
34% Independent = 202.32

Democrats: Obama 84%, Romney 10% = Obama 151.07, Romney 17.984
Republicans: Obama 5%, Romney 91% = Obama 8.992, Romney 163.65
Independents: Obama 36%, Romney 43% = Obama 72.835, Romney 86.998

Total: Obama 232.89, Romney 268.64 = Obama 41.44%, Romney 47.80%; Romney +6.36

Your efforts notwithstanding, and appreciated, it seems anticlimactic..

...it's all over but the shouting....

...and leaping from tall buildings by the Left.

It's way too early to start declaring victory. It's certainly starting to trend strong toward Romney (and Obama's statement on gay marriage just handed Romney North Carolina on a silver platter) but Romney has some problems in a few key states. A lot can happen.

I see our biggest problem being how President Romney performs...

 

[BluePhantom]

I see our biggest problem being how President Romney performs...

The current data certainly suggests that may be the case yes, but you know....something always happens. There will be an "October Surprise" of some sort (Obama's been laundering government money, Romney likes threesomes....you know, there's always something ) Plus the economy could dramatically recover, or tank even worse....who knows. But yes I agree with you (and the trend lines on the spreadsheet support) that Romney is on pace to win this.

 

[Foxfyre]

I just want to repeat my appreciation to Phantom for sharing his expertise in polling methods and statistical analysis and for keeping this thread supplied with some great material. Kudos!!!

And I agree that it's anybody's race to win or lose this far out from November. The Democratic attack machine is in full battle mode right now scouring every possible corner for ANYTHING that might look like dirt or can be made to look like dirt on Romney and they have a large surrogate media willing to put anything they find on the front page with a big headline or the lead story on the evening news. If we have polling organizations utilizing less than ethical methods to create news, that is definitely exacerbated by a media who doesn't give a flying fig about ethics.

And if they find anything really REALLY juicy, they will sit on it until October and release it just far enough ahead of the election so that the Romney camp will not have adequate time for damage control.