Home National Politics The Meaninglessness of Generic Polls

The Meaninglessness of Generic Polls


If you poll President Obama against a generic Republican, Obama trails 42.6% to 40.7%. So that means President Obama would trail any Republican opponent, right? That’s how the generic matchups are regularly treated by the political press.

But as Nate Silver shows at FiveThirtyEight.com, put Obama up against any actual Republican & he trounces ’em by as much as 14%:

  • Many of these polls are hopelessly simplistic.  Healthcare polls usually just ask if one supports the healthcare bill.  But when a few polls asked why, the story was very different.  Turns out a good number of those negative responses were from people who thought the healtcare bill was too weak.

  • Eric

    Here’s a timely one from CNN regarding how Americans feel with regard to how tax dollars are spent.


    The break out, as described in the article, isn’t as partisan as I would have thought with ~factoids~ like: “Republicans are split down the middle, with 49 percent calling it fair and 51 percent thinking it is unfair.“.  It’s moderately interesting at best but probably worth the minute it takes to read.

    Beyond that, not particularly useful as it doesn’t get into any sort of detail.  Asking people whether the govt wastes money or if taxes are fair is meaningless without getting into the “why?” details.

    • Elaine in Roanoke

      Polling during the debate on health reform always showed fairly strong support for the inclusion of a public option. I would love to see a poll on Rep. Alan Grayson’s idea of opening up Medicare to all who wanted it and would pay their actual cost for it.

      If people understood that we already have in place a mechanism for single payer universal medical care, it would be hard for the insurance lobby to fight it. Especially since the “problem” with funding for Medicare would disappear immediately.

  • Dems were sucked into similar sounding polls back in the mid 2000’s. Surely ANY Dem, the poll said, could take out Bush.  Good example of what looks one way on paper can be completely different when actual people and the real world is involved.  

  • Elaine in Roanoke

    I refuse to take any poll seriously that does not release a  breakdown of the polled universe by demographics, political leanings, etc.

    Plus, I refuse to take any poll seriously that does not release exactly how questions are worded. Far too often, the polling group does not understand how a question’s wording can unconsciously bias results obtained. Or, worse, the group DOES know how to bias results and uses that knowledge to attempt to give the results it desires (i.e., Faux News).  

  • kindler

    That’s Tim Pawlenty, right?