Rasmussen Reports Republican Bias?

Let’s take a look at Rasmussen’s final polls for the Senate races vs. the actual results to see if there is evidence of a systematic bias towards Republicans in their polling. I’m only looking at Senate races where there was a Rasmussen poll within the last month. There are some really obvious states (like Idaho & Vermont) where they didn’t bother polling in October.

State: Most Recent Rasmussen Poll vs. Actual Result = Difference

Hawaii: DEM +13 vs. DEM +54 = GOP +41
Nevada: GOP +4 vs. DEM +6 = GOP +10
Maryland: DEM +18 vs DEM +26 = GOP +8
Alaska: GOP +1 Vs IND +7 = GOP +8
California: DEM +3 vs. DEM +9 = GOP +6
West Virginia: DEM +4 vs. DEM +10 = GOP +6
Ohio: GOP +24 vs GOP +18 = GOP +6
Delaware: DEM +11 vs DEM +17 = GOP +6
Oregon: DEM +11 vs. DEM +16 = GOP +5
Colorado: GOP +4 vs. DEM +1 = GOP +5
New York (Schumer): DEM +28 vs DEM +33 = GOP +5
New York (Gillibrand): DEM + 21 vs. DEM +25 = GOP +4
Utah: GOP +33 vs GOP +29 = GOP +4
Conn.: DEM +7 vs DEM +10 = GOP +3
Indiana: GOP +18 vs GOP +15 = GOP +3
Pennsylvania: GOP +4 vs GOP +2 = GOP +2
Illinois: GOP +4 vs GOP +2 = GOP +2
Wisconsin: GOP +7 vs. GOP +5 = GOP +2
Washington: GOP +1 vs DEM +4 = GOP +5
Florida: GOP +20 vs. GOP +19 = GOP +1
Kentucky: GOP +12 vs. GOP +12 = Even
Arkansas: GOP +19 vs. GOP +21 = DEM +2
Missouri: GOP +9 vs GOP +13 = DEM +4
New Hampshire: GOP +17 vs. GOP +23 = DEM +6

Looking at this it seems quite obvious that Rasmussen has a significant bias towards the Republicans. Of these 24 polls, 20 had a lean towards the Republican, 1 was dead on, and 3 had a lean towards the Democratic candidate. The average lean was 4.9% points towards the Republican candidate.

In all 10 Senate elections decided by 10% points or less Rasmussen had a lean towards the Republican with the same average GOP bias of 4.9% points.

The Hawaii result is also interesting due to being off by such a huge amount. Was Rasmussen trying to to create a narrative of the Democrats struggling even in a deep blue state like Hawaii that happens to be the home state of the President? 538 says this 41% miss was the biggest one by any polling firm within the last 12 years.

I think it’s important to remember this bias during the next election cycle so that Rasmussen’s polling results include an asterisk for known GOP bias. I believe Rasmussen is in the business of narrative setting, and if their polls are known to be biased towards the GOP, they will be less successful in doing it.

I may have missed some October Senate polls, please let me know in the comments if you find any that I did not include.

AddThis Social Bookmark Button

5 Comment(s)

  1. Interesting, the only instances (3 of them) where Rassmussen predicted Dems would do better than they did, it was instances where Republicans were expected to win overwhelmingly and did so giving a few points to the Dems didn’t help them any. And in all 4 instances where the winner was different from their polling, Rassmussen chose the Republican Party candidate to win and they didn’t. There were no instances of Rassmussen predicting a Democratic Party candidate to win where the Republican or another party actually won. And in 19 of 23 races they predicted better Republican numbers than actually occurred. Even if they didn’t put their thumbs on the scales, it shows that their models favor the GOP more than the Democratic Party.

    They overestimated the Republican’s votes 83% of the time and they were 6.33 times as likely to overestimate Republican numbers than Democratic ones.

    More analysis should be done by those more proficient in statistics, but that’s quite a smoking gun.

    Joe | Nov 3, 2010 | Reply

  2. That 4.6% points in the 10 close Senate races looks like a smoking gun. That’s a nearly 5 point GOP bias with every single race having at least a 2 point bias.

    ThePresidentialCandidates.US | Nov 3, 2010 | Reply

  3. Also the Washington #s aren’t final. The Democrat will probably end up winning by more than +1 (because of where the rest of the ballots are.) So that will make the Rasmussen predictions even more off!

    ThePresidentialCandidates.US | Nov 3, 2010 | Reply

  4. I just realized that I missed Rasmussen’s biggest blown call of all: Hawaii!

    They had a poll out on October 17th ( http://www.rasmussenreports.com/public_content/politics/elections/election_2010/election_2010_senate_elections/hawaii/election_2010_hawaii_senate )

    showing the Democrat with 13% point lead (53% to 40%) …

    but the Democrat won the election by 54% points (75% to 21%)!

    Yes, they had a 41% point bias towards the Republican.


    ThePresidentialCandidates.US | Nov 6, 2010 | Reply

  5. I’m going to add Hawaii to the numbers and update Washington (more votes have been counted.)

    ThePresidentialCandidates.US | Nov 6, 2010 | Reply

Post a Comment

Plugin from the creators of Brindes :: More at Plulz Wordpress Plugins