Nate Silver Rates The Ubiquitous Presidential Polls


The end of election season means that we have at least a year (468 different Congressional races are just around the corner!) before we're again inundated by an endless stream of polls, proclamation and other numerical electoral predications.

Taking advantage of this lull, master stastician Nate Silver has rated the most prominent polls cited during the 2012 election. Gallup, one of the most popular, appears to be the least accurate.

An excerpt of Silver's analysis:

Among the more prolific polling firms, the most accurate by this
measure was TIPP, which conducted a national tracking poll for
Investors’ Business Daily. Relative to other national polls, their
results seemed to be Democratic-leaning at the time they were published.
However, it turned out that most polling firms underestimated Mr.
Obama’s performance, so those that had what had seemed to be
Democratic-leaning results were often closest to the final outcome.

Conversely, polls that were Republican-leaning relative to the consensus did especially poorly.

Well at least the GOP exceled at something this election season: failing.


  1. TonyJazz says

    I had always thought Gallup was one of the better ones. Were they just poorly executing their job or has the Gallup poll become another biased pro-Republican organization?

    Money corrupts…..

  2. Paul R says

    Clearly the Repubs polled a lot and lied about the results. Though I would never trust the results of online or robocall polls. Even live call and in-person polls are dubious, though slightly more reliable.

  3. jamal49 says

    Tony Jazz: That is the great mystery. Gallup used to be reliable as a polling company. But since at least 2004, they have slipped in their reliability and objectivity. Their polls are now automatically suspect. I wonder why.

  4. I wont grow up says

    Please say it isn’t so, not another endless election with lawn signs, radio ads, TV ads, rallies, pontificating political pollsters and all the rest arrrrrrrrggghhh!

    As citizens lets make a new law, they’re only allowed to campaign until 30 days before the election. If they don’t obey we strap them to a chair with headphones and make them listen to an endless loop of tibetan monks throat singing version of “Come on baby light my fire”.

  5. Moz's says

    gallup also rated near the bottom 2008 so gallup is done

    They did not learn their lesson from 2008 but heightened it. No reputable business or study group should ever hire Gallup except to clean out toilets

    & Ras? LOL no one who outside of fox audience accepts ras for anything more than what it is = republican propagandist

  6. Bill says

    Based on what he appears to have said (maybe he meant something a bit different), Nate Silver’s analysis is in this case somewhat off target. He seems to have measured errors and bias by comparing the poll results to the election results over a several week period. The problem with this is that it does not in any obvious way account for the possibility of public opinion changing during that time.

    Also, some polls may ask “how would you vote today,” which is the right question to ask if you are trying to evaluate how a political campaign’s marketing is doing. but a misleading question in the case where someone already mailed his or her vote in and you want to know how the election will turn out.

    Nate Silver’s definition of bias could be better as well: given that he ignored changes in public opinion over those weeks, a better number to estimate bias is to average all the polls from a given organization, and divide the difference of that from the actual results by the standard deviation of the average number.

  7. my2cents says

    does anyone here honestly believe that the repugs ever wanted to win this election?
    IMHO they wanted to win, and do their best-est to undermine any progress, bring the nation to its economic knees before going in for the jugular…

  8. anon says

    There is a well known effect of a sitting president getting +1 over his polling numbers on election day, so if you factor that into the numbers above (and we don’t know where they come from, really), then many of them were closer than they appear from this list, and many were well within their own margins of error.

Leave A Reply