Political Animal


May 09, 2011 1:45 PM Don’t poll about polls

By Steve Benen

I’ve read the report a few times, and as best as I can tell, this isn’t a joke.

Most voters believe the popularity boost President Obama has received following the killing of Osama bin Laden will disappear quickly, according to a new poll conducted for The Hill.

Almost three in five likely voters think the “bin Laden bounce” will vanish within three months, the poll indicates. Only one in four believes that Obama will still be benefitting in six months’ time.

So, let me get this straight. After President Obama announced the bin Laden news, news organizations put polls in the field and found his support had gone up. That makes sense — the media wanted to measure whether the killing of the al Qaeda leader affected the president’s standing, and it had.

But this is something different. The Hill conducted a poll asking people what they think might happen with other polls.

We’re in an environment in which we’re polling about Obama … and we’re polling about Obama polls.

Look, I enjoy going through survey data as much as the next political junkie, but this is deeply silly.

Steve Benen is a contributing writer to the Washington Monthly, joining the publication in August, 2008 as chief blogger for the Washington Monthly blog, Political Animal.


Post a comment
  • DAY on May 09, 2011 1:56 PM:

    We need to ask people if they think that people who answer polls are more or less informed on the questions asked.. Equal weight must be given to those who do not answer the poll. An anxious America awaits the results.

  • mikefromArlington on May 09, 2011 2:06 PM:

    Just reading the comments over there...

    "NY times and CBS poll…give me a damn break.The Jews and the libs and a poll that show a good result of the guy from Kenya…no doubtBY Paul on 05/09/2011 at 08:56"

    What a neat site. Since when is it OK for sites to allow racist to run rampant? The quality of their polling questions is reflected in the quality of their self policing of their site.

    What a waste of internet space.

  • jjm on May 09, 2011 2:07 PM:

    Wonderful post!

  • President Lindsay on May 09, 2011 2:09 PM:

    We need to add more questions to the poll and then report the correlation between those who view the president negatively and those who answer affirmatively to the following questions:

    1) Do you believe in angels?

    2) Do you believe in the devil?

    3) Did God create the earth about 6,000 years ago?

    4) Is teaching evolution in schools a terrible mistake?

    5) Should women be submissive to their husbands?

    Dare I say that the correlation would be strong, and in the 30% range? Why does anyone pay attention to polls from moronamericans?

  • john sherman on May 09, 2011 2:14 PM:

    Swift has a little poem that runs (quoting from memory):

    So naturalists observe a flea,
    With smaller fleas upon him prey,
    And others smaller yet to bite-em,
    And so proceed ad infinitum.

    It's bad natural history, but it well describes what often occurs in human affairs.

  • Matt on May 09, 2011 2:22 PM:

    Silly? Yes and no. It's silly if you take this at face value. But it does point out a very real phenomenon, which far too few pollsters take into account, which is that in most cases you cannot simply ask people questions and expect an answer that reflects their actual beliefs. Not anymore. Back in the early days of Gallup and Roper, or at least back before we all read five poll results a day with our morning coffee, maybe you could trust most people to give an "unstrategic" answer to the question you were asking. But you can't anymore.

    We see examples of this all the time. Liberal websites love to run hair-on-fire articles about the shockingly high percentage of Americans who reject evolution, or who support repeal of the First Amendment, or who think interracial marriage should be illegal--etc. But this simply doesn't reflect reality. What it does reflect is people's awareness that polls get published and talked about, and that there's political advantage to moving the window by gaming polls. Or simply that it's fun to take potshots in the culture war! If a pollster ever asks me for my party affiliation and then, having learned that I'm a Democrat, asks me how I feel about mandatory Sharia law (or a $20/gallon gas tax, or mandatory gay sex education in public schools, or some other conservative fever dream about what liberals believe), you bet I'll say I'm for those things. Let 'em sweat!

    Ask yourselves--if Fox News pollsters called you up under the mistaken impression that you were a New Hampshire Republican, and asked whether you'd support Mitt Romney or Michele Bachmann in the primaries, wouldn't you say you were a Bachmann backer all the way? Of course you would--just like a rock-ribbed Tea Partier might, with heavy heart, say he was for Romney. He's NOT--it's just that he knows Romney at least has a ghost of a chance against Obama, whereas Bachmann doesn't, so he's already thinking ahead to how his one answer will get magnified into a very heavy thumb on the scales as candidates make their decisions.

    Only very carefully worded polls on very carefully defined subjects can avoid this kind of strategic answering. Unfortunately, most polls don't bother, which in many cases is probably deliberate, the better to have "shocking" news to report.

  • paulo on May 09, 2011 2:30 PM:

    Like most trends in journalism it seems Monty Python was there first.

    As I recall, It was a sketch set on an island inhabited tv reporters and they all walked around interviewing each other.

  • ManOutOfTime on May 09, 2011 2:33 PM:

    I did an informal sampling of progressive blog commenters. seven of ten believe polls about polls are nonsense; one could not get his opinion posted because he could not decipher his Captcha; one is a concern troll who thinks it's another example of Obama getting a free ride from the MSM; and one believes THIS IS VERY GOOD NEWS!! FOR JOHN MCCAIN!!!!

  • chi res on May 09, 2011 2:44 PM:

    Not just a poll about a poll, mind you, but a poll asking people to prognosticate about the timing and results of a future poll based on a recent past poll.

    Who comes up with this shit? And how much do they pay? And will they hire me?

  • slappy magoo on May 09, 2011 3:54 PM:

    Despite attempts by the GOP, there's not much of a way to spin "he signed the order that killed Public Enemy No. 1" in a negative way. Obama is still fairly popular compared to many other Presidents at this stage of their first term, and when you see how weak the GOP field is for 2012, the media is looking for any way to make this a horse race. Can't spin getting bin Laden negatively? Spin the spin - HOW good will this be and for HOW long?

  • NancyPinSF on May 09, 2011 4:58 PM:

    Let me throw out a second opinion here. And I'll say up front that I may be giving The Hill too much credit. However, this could actually be a way to cut through the rah-rah and get at how people really feel about Obama in the wake of a very positive event for him politically. It might on the surface be a question about polls, but it also gives people a safe way to express mixed or negative feelings about Obama with a little distance -- sort of a check on the Bradley/Dinkins effect.

  • R on May 09, 2011 6:50 PM:

    Well, they could poll about polls, or they could do some actual investigative journalism, such as reading Ryan's bogus plan and doing the numbers...oh, wait. Never mind.

  • Rick Massimo on May 10, 2011 5:34 AM:

    Well, you have to understand: A poll came out with good news for a Democrat. It had to be balanced somehow, in this case by reassuring information that it's probably just a blip.