Scott Rasmussen Tries To Explain Why His 2012 Polling Was So Bad

Rassmussen Reports was among the least accurate pollsters of the 2012 Election Cycle, tying for 24th with Gallup out of 26 pollsters ranked in one post-election study. The main reason for this, of course, is the fact that Rasmussen went into his polls with a pre-conceived notion of what the electorate was supposed to look like and, based on those assumption, weighted his polls according to Party ID to fit into the model, something that practically no other pollster does. Four years ago, Rasmussen Reports was among the most accurate pollsters of the bunch, this year it what among the worst. So, what went wrong?

Scott Rasmussen tries to explain:

So what happened? “In general,” says Scott Rasmussen, “the projections were pretty good. The two differences I noted were share of white vote falling to 72 percent. That’s what the Obama campaign, to their credit, said all along. We showed it just over 73 percent. Also, youth turnout higher and senior turnout lower than expected. That’s a pretty big deal given the size of the generation gap. I think it showed clearly that the Obama team had a great game plan for identifying their vote and getting it to the polls.”

The problem with these polls—which are automated, as opposed to using live callers—was that they missed the correct model of the electorate. “The reality is that there were eight toss-up states,” says Rasmussen. “Some people projected Romney would do a couple of points better than the polls and sweep those states. Instead, it was Obama who did a bit better and swept them. I look at the campaign as about fundamentals. Obama job approval on Election Day was 50 percent. That meant there was a good chance he would get 50 percent of the vote. Also, 36 percent said their finances were in good shape. Up from 35 percent the day Obama took office. In other words, the fundamentals were just good enough for the president to keep his job.”

Interestingly, Rasmussen doesn’t address the flawed assumptions behind his own polling. Which makes me think he’s not going to correct his mistakes. Based on the performance of Rasmussen Reports in 2010 and this year, it’s hard to figure out why they should be taken seriously in the future absent a massive change to the way their polling is conducted.

FILED UNDER: 2012 Election, Public Opinion Polls, US Politics, ,
Doug Mataconis
About Doug Mataconis
Doug Mataconis held a B.A. in Political Science from Rutgers University and J.D. from George Mason University School of Law. He joined the staff of OTB in May 2010 and contributed a staggering 16,483 posts before his retirement in January 2020. He passed far too young in July 2021.

Comments

  1. EddieInCA says:

    Like many on the Wingnut Welfare Circuit, (Morris, Will, Barone, Rush, Hannity, Medved, Gallagher, Coulter, Levin, Malkin, Erickson, Trende, Cost, etc, etc) Rasmussen will continue to make gobs of money by ignoring reality and telling the GOP masses what they want to hear. This election, there was enough information, real information (not “guts”, “feelings”, “intuition”, “momentum”,) which forecasted which way the election was going to go. Almost to a person, the GOP “intellectuals” chose to ignore the facts.

    If 23 straight polls – from every single pollster, shows your guy losing by more than 3 points, he’s not going to miraculously win on election day. Period. Never has that happened before.

    But some people chose to ignore reality. The bigger question is why anyone takes these people seriously at all.

  2. Tsar Nicholas says:

    The main reason for this, of course, is the fact that Rasmussen went into his polls with a pre-conceived notion of what the electorate was supposed to look like and, based on those assumption, weighted his polls according to Party ID to fit into the model…..

    You’re actually giving him too much credit.

    The above would explain why his horserace calls were so wrong. But it doesn’t explain other elements of this polling. Like for example how he came up with a final call of R+5 for the total electorate but correspondingly a final call of Romney + 1 for the popular vote. That was oxymoronic and did not compute. That right there was a huge red flag and a major dead canary in the coal mine. If the electorate would have been R+5 then Romney would have prevailed in a huge landslide. Everyone after all knew Romney was going to win the Independent vote. Everyone knew Romney would obtain net crossover votes. So there was more going on here than what met the naked eye.

    Looking back it’s pretty clear that Rasmussen must have known all along Obama was ahead but decided nevertheless to sell a story to the political right. But he couldn’t very well give a final call of Romney 53-46. So he just pulled his final numbers out of his arse.

  3. OzarkHillbilly says:

    @EddieInCA:

    the GOP “intellectuals”

    Being awfully generous, aren’t you Eddie?

  4. An Interested Party says:

    Scott Rasmussen Tries To Explain Why His 2012 Polling Was So Bad

    …and fails, just like his polling…

  5. Aroostook says:

    The problem was he was funded by the Republicans and put that slant to it!

  6. jukeboxgrad says:

    Four years ago, Rasmussen Reports was among the most accurate pollsters of the bunch

    I wish I knew why people keep saying that. Rasmussen did a horrible job in 2008. For example, he predicted that FL would be McCain +1. Actual result: Obama +2.8. He also called NC for McCain, and Obama won that, too. He said OH was a tie. Actual result: Obama +4.6. He said NV would be Obama +4. Actual result: Obama +12.5. Et cetera. Link.

  7. sirous says:

    It’s amazing when Rasmussen still refers to his polls and says Obama job approval was 50%. He still seems didn’t get it. Your sampling was wrong and was not representing the target population.

  8. KariQ says:

    Four years ago, Rasmussen Reports was among the most accurate pollsters of the bunch

    No, Doug, he wasn’t. Please, please, go look it up. His national number was fine, but check out the states! He was wrong on every single swing state. Every. Single. One. And always too Republican.

    As I’ve said before, this is actually helpful for those of us who know to add +3 D to his numbers, but it still makes him less than accurate.

  9. jukeboxgrad says:

    those of us who know to add +3 D to his numbers

    In 2008 his median swing-state bias relative to the RCP average was 2.5, but his median bias relative to the actual result was 3.8. (That is, the RCP average also had a tilt, but it was smaller than his.)

    In 2012 his median swing-state bias relative to the actual result was 4.6. Link.

  10. Jr says:

    Scott shouldn’t be taking seriously after R+5. 2010 was only D+0, so why the hell would the 2012 electorate even be more white?

  11. Jo says:

    All they need to do to be more accurate is put the Rassmussen polling margin of error at +/- 10% instead of 3 or 4%. Whaalaa! Accurate.

  12. Geek, Esq. says:

    So he had the most diverse electorate ever part correct but also the most Republican electorate ever?

    Hack.

  13. Jim says:

    U would think that at the end Ramussen would realize bad results wont sell to anyone. Eventually even Repubs have got to realize buying pollsters does not buy election results. Then again, maybe the Repubs will just keep telling the same ole lies, there is no global warming, tax cut 4 the rich = jobs 4 the masses and skewing the polls in your favor actually changes the real results. Wake up! Telling lies doesn’t change the truth! Mitt Obama’s “gift” was honesty!