The difference between Rasmussen polling and non-Rasmussen polling has reached new heights in tonight's Senate forecast. For the first time, the non-Rasmussen forecast shows Democrats performing fully 3 seats better than in the forecast which includes Rasmussen polls.
Senate Picture, July 28th, with Rasmussen Most likely outcome: Democrats 52 seats, Republicans 47 seats, Charlie Crist one seat
Of the 100 Senate seats, 86 are either not up for re-election, or have a polling average where one party has a 100% chance of victory (if the election were held today). Among those 86 seats, there are 48 Democrats, and 38 Republicans. Here is a chart featuring the other 14 campaigns:
Senate picture, competitive campaigns chart, July 28th, with Rasmussen
The 48 currently safe Democrats, plus the 3.79 wins projected in these 14 campaigns, comes out to 51.79 Democrats, or 52 seats. Charlie Crist is also projected to win one seat.
Senate Picture, July 28th, without Rasmussen Most likely outcome: Democrats 55 seats, Republicans 44 seats, Charlie Crist one seat
Senate picture, competitive campaigns chart, July 28th, without Rasmussen
The 48 currently sage Democrats, plus the 6.69 wins projected in these 14 campaigns, comes out to 54.69 Democrats, or 55 seats. Charlie Crist is also projected to win one seat.
The "current Dem winning %" column projects the chance of Democratic victory if the election were held today. It is not meant to predict the chance of the Democratic candidate winning in November.
Every seat not listed here currently has either a 0% or a 100% chance of a Democratic victory.
Democrats are showing a distinct improvement in non-Rasmussen polling over the past two months, even as there has been little overall change within Rasmussen polling. They are now at 55 seats for the first time in over six months.
As the gap between Rasmussen and non-Rasmussen polls increases, I become more inclined to believe the non-Rasmussen forecast. Outlying polls, both by definition and by empirical observation, have more error than non-outlying polls. So, the further away Rasmussen moves from the average of all other polls, the more I think something is wrong with Rasmussen polling.
It wouldn't be the first time. In the past, Rasmussen has received a significant source of right-wing funding, and then proven to be the least accurate pollster in an election by erring massively in favor of Republicans.
The main story in today's Senate forecast is the re-emergence of a significant difference between Rasmussen and non-Rasmussen polls. Most strikingly, while Rasmussen has Republicans ahead (or safe / non up for re-election) in 50 Senate seats, the average of all other polls only shows Republicans ahead in 43 Senate seats. the seven differing states are Washington, Nevada, Colorado, Wisconsin, Ohio, Pennsylvania, and Florida.
Due to this difference, I have once again started producing forecasts with and without Rasmussen polls. Here are the numbers:
Senate Picture, July 26th, with Rasmussen Most likely outcome: Democrats 52 seats, Republicans 47 seats, Charlie Crist one seat
Odds of GOP taking chamber: 1.93% (if the election were held today)
Of the 100 Senate seats, 87 are either not up for re-election, or have a polling average where one party has a 100% chance of victory (if the election were held today). Among those 87 seats, there are 48 Democrats, and 39 Republicans. Here is a chart featuring the other 13 campaigns:
Senate picture, competitive campaigns chart, July 26th, with Rasmussen
The 48 currently safe Democrats, plus the 3.83 wins projected in these 13 campaigns, comes out to 51.83 Democrats, or 52 seats. Charlie Crist is also projected to win one seat.
Senate Picture, July 26th, without Rasmussen Most likely outcome: Democrats 54 seats, Republicans 45 seats, Charlie Crist one seat
Odds of GOP taking chamber: 0.02% (if the election were held today)
Senate picture, competitive campaigns chart, July 26th, without Rasmussen
The 48 currently sage Democrats, plus the 6.35 wins projected in these 13 campaigns, comes out to 54.35 Democrats, or 54 seats. Charlie Crist is also porjected to win one seat.
The "current Dem winning %" column projects the chance of Democratic victory if the election were held today. It is not meant to predict the chance of the Democratic candidate winning in November.
Every seat not listed here currently has either a 0% or a 100% chance of a Democratic victory.
Commentary on today's forecast can be found in the extended entry.
On Thursday, Chris wrote a diary, "Rasmussen vs non-Rasmussen Senate forecast" that further clarified the problem cluster of Rasmussen polling. After the table showing all the relevant races with and without Rasmussen polls included, Chris concluded:
Here are the projected Democratic seat totals for all three of these forecasts:
No Rasmussen: 54.37
Only Rasmussen: 51.69
All polls: 52.49
At first glance, Rasmussen appears to be weighing down the projection quite a bit. The overall forecast (52.49 Dems) is significantly closer to the Rasmussen-only projection (51.69 Dems) than to the non-Rasmussen projection (54.37 Dems).
For the 19 campaigns where there is a comparative trend, the Rasmussen poll average shows Republicans performing much better than the average of all other polls. The median pro-Republican house effect is 7.0%, and the mean is 6.4%. The difference is by no means consistent, as Rasmussen actually shows Democrats performing better in Pennsylvania and Washington than do other polls, and about the same as other polls in California, Iowa and Illinois. As such, just tacking on 6.4%, or 7.0%, to the Democratic candidate in Rasmussen polls would not likely increase the accuracy of an election forecast.
Add this to my observations in "Ramussen's junk polls distort House outlook", where I showed that Rasmussen's polls had a totally different trajectory over time than the average of other pollsters, and it becomes obvious that one simple, single explanation is not going fully explain what's going on with Rasmussen. But I do think it's important to understand as much as about them as possible. And in that spirit, I want to comment on an a late January release in which Rasmussen openly pushed a Tea-Bag friendly narrative of "mainstream/populism" vs. the "political elites":
How much is Rasmussen polling influencing the outlook on the 2010 Senate elections? To find out, I produced updated three forecasts this morning: one without any Rasmussen polls, one with only Rasmussen polls, and one with all polls. Here are the results:
The 22 Senate seats that might switch partisan control The chart below looks only at current primary campaign leaders. Campaigns where incumbent party currently leads by 18.5% or more are considered "safe," and not listed.
Methodology * = In Arizona, I am currently projecting J.D. Hayworth to be the Republican nominee
Here are the projected Democratic seat totals for all three of these forecasts:
No Rasmussen: 54.37
Only Rasmussen: 51.69
All polls: 52.49
At first glance, Rasmussen appears to be weighing down the projection quite a bit. The overall forecast (52.49 Dems) is significantly closer to the Rasmussen-only projection (51.69 Dems) than to the non-Rasmussen projection (54.37 Dems).
For the 19 campaigns where there is a comparative trend, the Rasmussen poll average shows Republicans performing much better than the average of all other polls. The median pro-Republican house effect is 7.0%, and the mean is 6.4%. The difference is by no means consistent, as Rasmussen actually shows Democrats performing better in Pennsylvania and Washington than do other polls, and about the same as other polls in California, Iowa and Illinois. As such, just tacking on 6.4%, or 7.0%, to the Democratic candidate in Rasmussen polls would not likely increase the accuracy of an election forecast.
Personally, I am still going to stick with the average of all polls, for now. In the meantime, I will be conducting some research to see if weighting polls by "house effect" would have produced more accurate forecasting results in the 2004, 2006 and 2008 elections. If it does, then I will start weighting polls by house effects. If it does not, then I won't.
Gallup and Rasmussen have (finally) started releasing their weekly tracking poll results on Monday, instead of Tuesday, allowing for a generic ballot update today.
Generic Congressional Ballot, May 24th: Democrats +0.5% Polls from here and here
This is good news for Democrats, but also only represents a swing of only 2.0% in their favor from the largest Republican advantage in my averages. This is really is only a slight improvement, not a dramatic one.
It is also worth noting that Rasmussen generates an overall swing of 1.41% in favor of Republicans. Without Rasmussen, Democrats lead by 1.95% in my average, pushing the average to Dem +0.54% from Dem +1.95%. Pollster.com shows a 0.8% swing in favor of Democrats without Rasmussen.
I don't really know what to do with Rasmussen polling at this point. Rasmussen is clearly outlying from other polling, which is a big problem given that it produces a huge percentage of polling public polling on Senate, Governor and House campaigns. One thing I am not going to do is tweak Rasmussen polling, by leaving it in the averages but adding 6% (or so) to the Democratic candidate. Either I will continue to include Rasmussen in the averages with no tweaking, or it will just be dumped from the averages altogether. The best solution is probably just to produce polling averages both with and without Rasmussen numbers, with the former still receiving the topline credit.
A new study from the CDC shows that 24.5% of US households, and 22.9% of US adults, were wireless-only in the second half of 2009. This is a sharp increase over the past two years. In late 2007, 14.7% of US households, and 13.6% of US adults, were wireless-only.
This rapidly increasing level of wireless-only Americans will have an impact on political polling. Specifically, excluding wireless-only adults from political surveys will have a statistically significant, negative impact on Democratic performance in political polling. This was confirmed in a recent study by Pew, which compared the national generic ballot preference of a landline-only sample of 4,683 registered voters with a combined landline and cell-phone sample of 7,055 registered voters:
In the landline sample, Republican candidates have a 47%-to-41% margin over Democratic candidates on the 2010 generic horserace, but in the combined sample voters are evenly divided in their candidate preferences for this November (44% for each party).
There is still a margin of error in a poll with such a large sample size, but it is just barely over plus or minus 1%. As such, with an overall six-point gap, the survey shows a statistically significant difference between polls that include cell-phone only adults and polls that do not.
Previous research has shown that Rasmussen's use of a likely voter screen is not the reason why their polls now differ strongly from the trendline of all other polls. Rasmussen's likely voter polls are about six points more favorable to Republicans than other likely voter polls. Also, Rasmussen polls of all adults are six points more favorable to Republicans than other polls of all adults. This six-point pro-Republican tilt is exactly the gap found by Pew in their landline-only sample.
Since Rasmussen Reports excludes wireless-only adults from their surveys (possibly due to restrictions on automated phone calls to cell phones), it is likely that the wireless-only effect is one of the main reasons that Rasmussen surveys are now outlying compared to other polling outfits. Further, the lack of wireless-only adults also plays a role in the recent, particularly eye-popping Rasmussen results in Kentucky (Rand Paul up by 25%) and Arkansas (Boozman up by 27% on Bill Halter). In 2007, state-level data showed that cell-phone only adults were particularly prevalent in Arkansas and Kentucky. While 13.6% of the nation as a whole was cell-phone only in 2007, in Arkansas that figure was 21.2% and in Kentucky it was 21.6% (PDF, page 5). Given the trends nationwide, by May 2010 it is likely that roughly one in three adults in Arkansas and Kentucky are wireless-only. This would make for an even more pronounced, localized gap than the national discrepancy found in Pew's 2010 study.
The rapidly rise in wireless-only adults, along with the confirmation of the pro-Democratic tilt of those adults, helps explain some, and possibly all, of the difference between Rasmussen polls and other polling firms. Americans are dumping landlines at a rapid rate, and those Americans doing do skew heavily Democratic.
Update: As an election forecaster, I honestly don't know what to do about the emerging gap between polls that include cell-phone only and those that do not. About all I can say is that "it's a problem." I strongly prefer to base my methodology on empirical studies of what has worked in the past when it comes to poll-based electoral forecasting, but the emergence of a very sizable, heavily-skewed cell-phone only population is a relatively new problem. Polls from 2004-2008 will not offer much guidance here.
KY-Sen: Rand Paul flips on discrimination in private hiring: A spokesperson for Rand Paul has confirmed to Greg Sargent that Paul thinks the federal government should ban private businesses from discriminating on the basis of race in their hiring practices. This is the exact opposite of what Paul has stated in the past, not what he has said in the past, and just last night he said that the federal government did not have the right to prevent private businesses from refusing to serve people on the basis of race. So much for Rand Paul at least being principled.
Meanwhile, Rasmussen shows Rand Paul ahead by 25%. Their polling in Kentucky has been particularly henke, to the tune of a 14% house effect so far this year. Non-Rasmussen polling in Kentucky current shows Paul ahead by only 1%
NV-Sen: Angle takes the lead among Republicans: A new poll shows wingnut fave Sharron Angle ahead in the Republican primary for Nevada Senate. Tomorrow, a Rasmussen poll will show her beating Harry Reid by 22 points in the general election.
PA-Sen: Sestak takes the lead: The first post-primary poll in Pennsylvania shows Joe Sestak ahead of Pat Toomey, 46%-42%. And its from Rasmussen. Don't get too excited though, because next week Rasmussen will show Toomey ahead by 18%.
Also, according to Rasmussen polling, you trail your Republican neighbor by 37%, even though you have higher name ID in the neighborhood. Other polls show it to be a toss-up. Discuss.
Some new data suggests Americans hold a negative opinion about the state of the country and are optimistic as well. How can this be? At least two explanations are possible and there's data for each:
1. People are unhappy about how things are going now but hopeful about the future.
and/or
2. The country is split. Some optimists, some pessimists.
The latter split, according to one report, is not along party lines but a split between populists and the "ruling class." Details ahead.
On March 17, NPR published a poll under the innocuous headline "More Voters Think U.S. Is On Right Track." That poll had a rather surprising result on the question of generic Congressional ballot preference, with Democrats and Republicans tied at 42, and 15% undecided. Whoa. Is that true?
Well if you check the pollster trendlines, this seems entirely plausible:
I just want to do a brief review here of some of the recent signs pointing to an ever-more-likely landslide victory scenario this November, with special emphasis on the House.
I'm combining information from three sources: Rasmussen, Democracy Corps and Cook Political Report.
First, there's the latest information on Partisan ID from Rasmussen, based on their daily tracking polls. Here's the chart I generated from their data. It shows how Democratic Party identification jumped sharply from the beginning of the primary season, after falling into the doldrums during 2007, when the Democratic Congress basically failed to deliver much of anything. It's remained fairly steady the last few months, but at record high levels:
A bit explanation and the table the chart is based on can be found on the flip, along the goods from Democracy Corps and Cook Political Report.
Cross posted on from DailyKos with a different poll
Rasmussen recently published results of their first (as far as I know) poll involving more than the two major party candidates. While I am glad to see some reality based polling going on (there will be 5 or 6 candidates on most of the ballot for Pres in most states in November), Rasmussen has some pretty serious flaws in methodology (shocking, I know).
Obama earns 42% of the vote, McCain 38%, Bob Barr 6% and Ralph Nader 4%. Given those options, 11% were undecided.Barr and Nader were mentioned as candidates of the Libertarian Party and the Green Party respectively.
By comparison, Sunday's 2-way numbers were Obama 45 McCain 44 - clearly this is good news for Obama, even if the real pull the lever November support of these two candidates is only half what is shown here. Details on the poll numbers, as well as analysis of both the results and Rasmussen's flawed methodology after the jump.
First the Republicans began losing voters, then, only quite recently the Democrats began gaining them. Although the magnitude is less than that shown in a recent Pew Poll, the broad storyline is the same. Indeed, within the last month, the Rasmussen figures have suddenly jumped into the same range that Pew gave:
STATE BLOGS
