BluePhantom
Educator (of liberals)
- Thread starter
- #221
Weekly Breakdown for 6/12/12
We finally had a much better week this time and there is some good data to crunch. That’s not to say that there were not some things that annoyed me to death and so this week’s “Blue Middle Finger Award” once again goes to TIPPonline. Why? Well, despite the fact that they give some decent crosstabs to work with (much more than others actually), their methodology is starting to show a pattern of bias. They release a poll every other week and in each one, Obama’s numbers are right there with the rest of the professionals, but Romney’s numbers are always significantly lower. For the week ending 5/8 they had Romney 4.66 below the average of other pollsters. For the week ending 5/22 it was 5.66 points and this week it was a little better at 3.66. That’s an average of 4.66 points below the other professionals combined. Each week that they have released a poll they have the lowest Romney performance of any of the professional polls released that week and the biggest Obama spread. I don’t know why this is because they don’t release a full report of their questioning methodology for me to investigate, but a pattern of bias is developing and it’s really fucking up the trendlines.
The problem of course is that we can’t simply toss out the TIPP polls in our calculations because that would bias the entire process and open the door for unsavory manipulation of the data, so we’re forced to go with what we have, however it’s worth noting that if you toss out the TIPP polls simply for the sake of argument, Obama’s trend has flatlined while Romney’s is falling, but not nearly as much as with the TIPP polls. I don’t mind an occasional outlier or obviously biased poll every now and then, but when it’s a consistent pattern that occurs regularly enough that it begins to significantly influence the results I begin to get annoyed. So a big “fuck you” to TIPP this week.
This is also the first week where the media polls actually show a closer race than the professional polls, but again that’s because of the TIPP poll. Without that the media polls show better Obama performance as would be expected.
With that said let's look at the statistics:
6/5/12 - 6/12/12
Base Average
Professional Polls
Rasmussen (LV - tracking): Obama 45, Romney 46
Gallup (RV - tracking): Obama 46, Romney 45
Monmouth/SUSA/Braun (LV): Obama 47, Romney 46
IBD/CSMTIPP (RV): Obama 46, Romney 42
Professional Average: Obama 46.00, Romney 44.75; Obama +1.25
Media Polls
Reuters/Ipsos (RV): Obama 45, Romney 44
FOXNews (RV): Obama 43, Romney 43
Media Average: Obama 44.00, Romney 43.50; Obama +0.50
Combined Base Average: Obama 45.33, Romney 44.33; Obama +1.00
Adjusted Average
Professional Polls
Rasmussen (LV - tracking): Obama 675, Romney 690
Gallup (RV - tracking): Obama 1403, Romney 1372.50
Monmouth/SUSA/Braun (LV): Obama 541.44, Romney 521.92
IBD/CSMTIPP (RV): Obama 386.86, Romney 353.22
Total: Obama 3006.3, Romney 2945.6
Professional Average: Obama 45.95, Romney 45.02; Obama +0.93
Media Polls
Reuters/Ipsos (RV): Obama 381.6, Romney 373.12
FOXNews (RV): Obama 390.01, Romney 390.01
Total: Obama 771.6, Romney 763.13
Media Average: Obama 43.97, Romney 43.48; Obama +0.48
Combined Adjusted Average: Obama 45.53, Romney 44.69; Obama +0.83
Two Week Rolling Average
Base Average
Professional Polls
Obama 46.29, Romney 44.57; Obama +1.71
Media Polls
Obama 45.67, Romney 44.33; Obama +1.33
Combined Average
Obama 46.10, Romney 44.5; Obama +1.60
Adjusted Average
Professional Polls
Obama 46.39, Romney 44.59; Obama +1.80
Media Polls
Obama 45.67, Romney 44.33; Obama +1.33
Combined Average
Obama 46.27, Romney 44.55; Obama +1.73
In state polls I largely discussed them in my post yesterday so I will be brief here.
In Colorado we have two polls, both of which are LV. Rasmussen showed a 45/45 split while Purple Strategies showed a 48/46 Obama advantage. Both are statistical ties obviously and combined with last week’s NBC RV poll showing a one point Obama advantage we can now safely declare that one of two things is happening in Colorado and both of them are bad for Obama. A) Things are much tighter in the state than previously thought but there has been no real trend movement because the ProjectNew America/Keating(D) poll a couple weeks ago was total bullshit, or B) Things are much tighter in the state than previously thought and there has been heavy trending toward Romney recently. My gut tells me it’s A, but for statistical purposes we must assume it’s B. Future polls will reveal the truth.
In Florida a Purple Strategies LV poll showed a 45/49 Romney advantage. This defies the PPP(D) and NBC polls of the last two weeks that each showed four point Obama leads and supports the most recent Quinnapiac poll showing a six point Romney advantage for the week ending 5/22.
Rasmussen released an LV poll of Iowa showing a 46/47 Romney advantage. This supports last week’s NBC poll showing a 44/44 tie and if the PPP(D) poll for the week ending 5/8 is to be believed it indicates a very hard Romney trend in Iowa right now.
Rasmussen published an LV poll of Missouri showing a 42/49 Romney advantage. Last week PPP(D) had Obama up by a point in an RV poll. If we adjust for PPP’s historical house bias of around four point and adjust for the RV vs. LV effect of about five points it would indicate pretty much the same thing. Missouri is quickly becoming a lost cause for team Obama.
A PPP(D) RV poll of North Carolina revealed Obama trailing Romney by two 46/48. This is significant because even with PPP’s house bias they show Romney leading in the state. The trendline for Romney is off the charts in North Carolina and rises almost two points above the extended average. This suggests a massive momentum shift for Romney in North Carolina. Like Missouri, something very significant will have to happen for Obama to challenge in the state.
A Purple Strategies LV poll of Ohio showed a 45/48 Romney advantage. While this is within the MOE and is a statistical tie, it’s further evidence that Ohio is starting to move to the Romney camp. This poll supports last week’s Rasmussen results showing a two point Romney lead in the Buckeye State. As noted yesterday Obama’s trend has flatlined in Ohio while Romney’s rises slightly above the extended average. Projected toward election day it remains a statistical tie with a Romney lean.
In Pennsylvania we have two polls. One is a Quinnapiac RV poll showing a 46/40 Obama advantage and the other is a Franklin & Marshall RV poll showing a 48/36 Obama advantage. Now while I will freely admit that Pennsylvania is pretty clearly leaning toward Obama, a 12 point split is a little hard for me to fathom. Personally I will go more with Quinnapiac’s results but for statistical purposes we have to calculate both of them in. At the end of the day it really doesn’t matter all that much because even if we removed the F&M poll there would only be a slight increase in Romney’s trendline. As it is both trendlines are flat as a board with about an eight point Obama advantage. Even if we adjust for the RV vs. LV effect it would bring it to about a three point Obama advantage. Like Missouri and North Carolina for Romney, this is Obama’s state to lose right now and the trendlines show very stubborn resistance to current events that may shift the trends.
In Virginia we have two polls as well. One is a Quinnapiac RV poll showing a 47/42 Obama advantage and the other is a Purple Strategies LV poll showing it 49/46 in Obama’s favor. Adjusting for RV/LV basically what this suggests is a statistical tie in the state with an Obama lean. That would fit in line with what everything else suggests but it’s the trends that matter and right now Obama’s has flatlined like it has in other states while Romney’s is rising above the extended average…a little. It suggests that Romney may close the gap but it’s still going to be very close and while a toss up, it’s still leaning ever so slightly toward Obama. This is the one shining crown jewel of good news for the Obama camp this week amid a sea of disasters. Without Virginia it becomes very difficult for Romney to win this election. Not impossible mind you, but he would have to take Colorado and New Hampshire and either Nevada or Iowa. Virginia is critical and right now Obama is holding on by his toenails.
Finally in Wisconsin a WeAskAmerica LV poll shows a 48/43 Obama advantage. What I find amusing about this is that when you go to their website and click the link where they are supposed to be bragging about their accuracy it takes you to an independent agency that essentially says they suck. I was going to give them a Blue Middle Finger Award for that alone but I felt sorry for them. Poor little devils. Anyhow, it’s probably correct that Wisconsin is leaning toward Obama. There’s no surprise there and five points really isn’t out of the realm of reality so we’ll go with it but my gut feeling is that it’s probably more like two or three.
Did I say I was going to be brief? So much for that idea.
We finally had a much better week this time and there is some good data to crunch. That’s not to say that there were not some things that annoyed me to death and so this week’s “Blue Middle Finger Award” once again goes to TIPPonline. Why? Well, despite the fact that they give some decent crosstabs to work with (much more than others actually), their methodology is starting to show a pattern of bias. They release a poll every other week and in each one, Obama’s numbers are right there with the rest of the professionals, but Romney’s numbers are always significantly lower. For the week ending 5/8 they had Romney 4.66 below the average of other pollsters. For the week ending 5/22 it was 5.66 points and this week it was a little better at 3.66. That’s an average of 4.66 points below the other professionals combined. Each week that they have released a poll they have the lowest Romney performance of any of the professional polls released that week and the biggest Obama spread. I don’t know why this is because they don’t release a full report of their questioning methodology for me to investigate, but a pattern of bias is developing and it’s really fucking up the trendlines.
The problem of course is that we can’t simply toss out the TIPP polls in our calculations because that would bias the entire process and open the door for unsavory manipulation of the data, so we’re forced to go with what we have, however it’s worth noting that if you toss out the TIPP polls simply for the sake of argument, Obama’s trend has flatlined while Romney’s is falling, but not nearly as much as with the TIPP polls. I don’t mind an occasional outlier or obviously biased poll every now and then, but when it’s a consistent pattern that occurs regularly enough that it begins to significantly influence the results I begin to get annoyed. So a big “fuck you” to TIPP this week.
This is also the first week where the media polls actually show a closer race than the professional polls, but again that’s because of the TIPP poll. Without that the media polls show better Obama performance as would be expected.
With that said let's look at the statistics:
6/5/12 - 6/12/12
Base Average
Professional Polls
Rasmussen (LV - tracking): Obama 45, Romney 46
Gallup (RV - tracking): Obama 46, Romney 45
Monmouth/SUSA/Braun (LV): Obama 47, Romney 46
IBD/CSMTIPP (RV): Obama 46, Romney 42
Professional Average: Obama 46.00, Romney 44.75; Obama +1.25
Media Polls
Reuters/Ipsos (RV): Obama 45, Romney 44
FOXNews (RV): Obama 43, Romney 43
Media Average: Obama 44.00, Romney 43.50; Obama +0.50
Combined Base Average: Obama 45.33, Romney 44.33; Obama +1.00
Adjusted Average
Professional Polls
Rasmussen (LV - tracking): Obama 675, Romney 690
Gallup (RV - tracking): Obama 1403, Romney 1372.50
Monmouth/SUSA/Braun (LV): Obama 541.44, Romney 521.92
IBD/CSMTIPP (RV): Obama 386.86, Romney 353.22
Total: Obama 3006.3, Romney 2945.6
Professional Average: Obama 45.95, Romney 45.02; Obama +0.93
Media Polls
Reuters/Ipsos (RV): Obama 381.6, Romney 373.12
FOXNews (RV): Obama 390.01, Romney 390.01
Total: Obama 771.6, Romney 763.13
Media Average: Obama 43.97, Romney 43.48; Obama +0.48
Combined Adjusted Average: Obama 45.53, Romney 44.69; Obama +0.83
Two Week Rolling Average
Base Average
Professional Polls
Obama 46.29, Romney 44.57; Obama +1.71
Media Polls
Obama 45.67, Romney 44.33; Obama +1.33
Combined Average
Obama 46.10, Romney 44.5; Obama +1.60
Adjusted Average
Professional Polls
Obama 46.39, Romney 44.59; Obama +1.80
Media Polls
Obama 45.67, Romney 44.33; Obama +1.33
Combined Average
Obama 46.27, Romney 44.55; Obama +1.73
In state polls I largely discussed them in my post yesterday so I will be brief here.
In Colorado we have two polls, both of which are LV. Rasmussen showed a 45/45 split while Purple Strategies showed a 48/46 Obama advantage. Both are statistical ties obviously and combined with last week’s NBC RV poll showing a one point Obama advantage we can now safely declare that one of two things is happening in Colorado and both of them are bad for Obama. A) Things are much tighter in the state than previously thought but there has been no real trend movement because the ProjectNew America/Keating(D) poll a couple weeks ago was total bullshit, or B) Things are much tighter in the state than previously thought and there has been heavy trending toward Romney recently. My gut tells me it’s A, but for statistical purposes we must assume it’s B. Future polls will reveal the truth.
In Florida a Purple Strategies LV poll showed a 45/49 Romney advantage. This defies the PPP(D) and NBC polls of the last two weeks that each showed four point Obama leads and supports the most recent Quinnapiac poll showing a six point Romney advantage for the week ending 5/22.
Rasmussen released an LV poll of Iowa showing a 46/47 Romney advantage. This supports last week’s NBC poll showing a 44/44 tie and if the PPP(D) poll for the week ending 5/8 is to be believed it indicates a very hard Romney trend in Iowa right now.
Rasmussen published an LV poll of Missouri showing a 42/49 Romney advantage. Last week PPP(D) had Obama up by a point in an RV poll. If we adjust for PPP’s historical house bias of around four point and adjust for the RV vs. LV effect of about five points it would indicate pretty much the same thing. Missouri is quickly becoming a lost cause for team Obama.
A PPP(D) RV poll of North Carolina revealed Obama trailing Romney by two 46/48. This is significant because even with PPP’s house bias they show Romney leading in the state. The trendline for Romney is off the charts in North Carolina and rises almost two points above the extended average. This suggests a massive momentum shift for Romney in North Carolina. Like Missouri, something very significant will have to happen for Obama to challenge in the state.
A Purple Strategies LV poll of Ohio showed a 45/48 Romney advantage. While this is within the MOE and is a statistical tie, it’s further evidence that Ohio is starting to move to the Romney camp. This poll supports last week’s Rasmussen results showing a two point Romney lead in the Buckeye State. As noted yesterday Obama’s trend has flatlined in Ohio while Romney’s rises slightly above the extended average. Projected toward election day it remains a statistical tie with a Romney lean.
In Pennsylvania we have two polls. One is a Quinnapiac RV poll showing a 46/40 Obama advantage and the other is a Franklin & Marshall RV poll showing a 48/36 Obama advantage. Now while I will freely admit that Pennsylvania is pretty clearly leaning toward Obama, a 12 point split is a little hard for me to fathom. Personally I will go more with Quinnapiac’s results but for statistical purposes we have to calculate both of them in. At the end of the day it really doesn’t matter all that much because even if we removed the F&M poll there would only be a slight increase in Romney’s trendline. As it is both trendlines are flat as a board with about an eight point Obama advantage. Even if we adjust for the RV vs. LV effect it would bring it to about a three point Obama advantage. Like Missouri and North Carolina for Romney, this is Obama’s state to lose right now and the trendlines show very stubborn resistance to current events that may shift the trends.
In Virginia we have two polls as well. One is a Quinnapiac RV poll showing a 47/42 Obama advantage and the other is a Purple Strategies LV poll showing it 49/46 in Obama’s favor. Adjusting for RV/LV basically what this suggests is a statistical tie in the state with an Obama lean. That would fit in line with what everything else suggests but it’s the trends that matter and right now Obama’s has flatlined like it has in other states while Romney’s is rising above the extended average…a little. It suggests that Romney may close the gap but it’s still going to be very close and while a toss up, it’s still leaning ever so slightly toward Obama. This is the one shining crown jewel of good news for the Obama camp this week amid a sea of disasters. Without Virginia it becomes very difficult for Romney to win this election. Not impossible mind you, but he would have to take Colorado and New Hampshire and either Nevada or Iowa. Virginia is critical and right now Obama is holding on by his toenails.
Finally in Wisconsin a WeAskAmerica LV poll shows a 48/43 Obama advantage. What I find amusing about this is that when you go to their website and click the link where they are supposed to be bragging about their accuracy it takes you to an independent agency that essentially says they suck. I was going to give them a Blue Middle Finger Award for that alone but I felt sorry for them. Poor little devils. Anyhow, it’s probably correct that Wisconsin is leaning toward Obama. There’s no surprise there and five points really isn’t out of the realm of reality so we’ll go with it but my gut feeling is that it’s probably more like two or three.
Did I say I was going to be brief? So much for that idea.
Last edited: