The Bishop and the Butterfly: Murder, Politics, and the End of the Jazz Age
    Donal's picture

    Rasmussen Skews Red

    By Donal on Fri, 11/05/2010 - 5:28am | Politics



    Nate Silver grades the pollsters and finds Rasmussen biased.

    On Tuesday, polls conducted by the firm Rasmussen Reports — which released more than 100 surveys in the final three weeks of the campaign, including some commissioned under a subsidiary on behalf of Fox News — badly missed the margin in many states, and also exhibited a considerable bias toward Republican candidates.

    Other polling firms, like SurveyUSA and Quinnipiac Univesrity, produced more reliable results in Senate and gubernatorial races. A firm that conducts surveys by Internet, YouGov, also performed relatively well.
    ...

    Other polling firms that joined Rasmussen toward the bottom of the chart were Marist College, whose polls also had a notable Republican bias, and CNN/Opinion Research, whose polls missed by almost 5 points on average. Their scores are less statistically meaningful than that for Rasmussen Reports, however, because they had only released surveys in 14 and 17 races, respectively, as compared to Rasmussen’s 105 polls.



    In another piece, Silver fingers a lack of consideration for the Hispanic or Latino vote for most pollsters getting the Reid and Bennett picks wrong.

    There is another theory, however, which was proposed to me last night by Matt Barreto of the polling firm Latino Decisions.

    “There is one overarching reason why the polls were wrong in Nevada,” Mr. Barreto wrote in an e-mail to FiveThirtyEight. “The Latino vote.”

    His firm, which conducts interviews in both English and Spanish, had found that Latino voters — somewhat against the conventional wisdom — were relatively engaged by this election and for the most part were going to vote Democratic. Mr. Barreto also found that Latino voters who prefer to speak Spanish — about 40 percent of Latino voters in California meet this description, he told me — are particularly likely to vote Democratic. Pollsters who don’t conduct bilingual interviewing at all, or who make it cumbersome for the respondent to take the poll in Spanish, may be missing these voters.


    So when you're a Jet, you're a Jet all the way, but you should know which way the Sharks are voting, too.

     

    Topics: 
    Politics

    Comments

    When I read the title of the post, my first thought was, "In other news, grass is green..." But I'm glad somebody is analyzing this stuff, even if it makes my head hurt a little bit to think about it. Because Rasmussen is so far away from the elusive zero, I have to think they are doing it on purpose. Possibly that's because Republican candidates will pay for the good news and it's financially lucrative to Rasmussen as a polling company. I can't think of any other payoffs. Most media outlets commission their own polls during election years, so the Rasmussen polls aren't getting any wider play than others, especially on the major networks. Yes, people want to back winners, but what kind of psychological impact does a good poll truly have on a block of voters? I don't know. I can't see the upside. Maybe somebody can explain it to me.


    by Orlando on Fri, 11/05/2010 - 7:22am

    Joyner suggests that they're cheap rather than biased:

     

    Rasmussen’s sample is biased because they’re polling on the cheap — using robocalls, which by law can’t dial cell phones, and otherwise cutting corners — rather than because of some agenda to propagandize for the GOP.   The end result, however, is the same: Polls that can’t be trusted.

     

    Could be both, of course.


    by Donal on Fri, 11/05/2010 - 12:42pm

    Ok, maybe. But if they know they're doing it, isn't there some mathematical model they could apply to correct for the difference? 


    by Orlando on Fri, 11/05/2010 - 2:27pm

    Hard to say. Suppose I balance my checkbook. I never get a receipt from the parking garage ~$8 . I often fail to get a receipt from the gas pump ~$30, and often I'm unwilling to go into the station and wait for them to print one. My wife always writes a few checks for this and that and forgets to put them in the register ~$20 - ~$50. So when I balance my checkbook, I only input the easy amounts for which I have receipts or register entries. What mathematical model should I use to correct for the difference?


    by Donal on Fri, 11/05/2010 - 3:02pm

    They could, as many pollsters do, model the going rate of their own bias, and try to compensate for it. Or they could try to fix their sample by calling a larger overall number of respondents and throwing out more of the responses that are currently oversampled. If they're getting a population that is 20% whiter than the actual voting population, or has 25% more senior citizens, or is 15% more likely to self-identify as Republicans, you can work out a proportion that uses fewer of those responses. Or they could go crazy and use actual employees instead of auto-dialers and computers.

    The point that they're cheap seems to be the valid one, or an overlap between a (procedural) bias that comes from corner-cutting and an (ideological) bias that likes the resulting Republican lean and therefore doesn't want to question it.  Of course, in recent years cutting corners to keep your labor costs down, quality be damned, has turned into a core Republican value.

    What's most surprising, or most unsurprising, is that Rasmussen is stonewalling rather than responding to Silver's analysis. That he wants to stick with his model even after its results have started to be unreliable suggests a truly Bushlike world view.


    by Doctor Cleveland on Mon, 11/08/2010 - 1:20pm