Thursday, February 10, 2011

Is There Any Info In Early Polls and Straw Polls?

Brendan Nyhan cautions everyone today to pay little attention to the straw poll results at CPAC or to early horse race polling for the GOP presidential nomination:
At this point in the election cycle, the preferences that matter are those of the activists, elected officials, donors, and party elites who take part in the so-called "invisible primary." Media hype of public opinion surveys and straw polls only serves to obscure where the meaningful action is taking place.
I agree with him that the front-line numbers for both of these indicators are very easy to overhype. I'm not sure they are entirely meaningless, however.

For example, if Sarah Palin was in fact polling as a strong frontrunner right now -- say, getting around 40% of the vote in horse race matchups -- I'd be taking her a lot more seriously as a candidate than I am. More to the point, "activists, elected officials, donors and party elites" would be taking her a lot more seriously.

In general, I'd think that polling is meaningful in more or less the way that primaries were meaningful in the old nomination system before reform (that is, pre-1972). They're one piece of evidence that people who do have a major say in the nomination may look at when trying to choose which candidate to back. Now, they -- and we -- need to be very careful about how to interpret that evidence, since name recognition, and not actual preference, is going to be the main thing that these polls measure. I'm not convinced, however, that there's really no value added beyond that.

Straw polls are a somewhat different story. The voters are the kinds of people who matter during the invisible primary, so one might think that a public tally of where some of them stand would hint at what's happening in general. Unfortunately, most straw polls are able to be manipulated by the candidates (for example, by busing in supporters to an event), so they may tell us less about what activists think than which activists happened to be in the room. On the other hand, any candidate who tries hard to organize a strong vote in an early straw poll and fails miserably may be letting us know that there's something wrong.

For example, in 2007 Mike Huckabee did badly at CPAC. I have no idea whether or not he tried to compete in that straw poll, but if he did, it might have been an indication of the hostility of some movement conservatives to his candidacy -- the same ones who wound up derailing his campaign after his Iowa breakthrough in 2008. So it seems to me that straw polls might hold some useful hints about what's going on, but they can only be interpreted with the help of a good deal of on-the-ground reporting.

So, overall I agree with Nyhan, but I'd also try to squeeze a bit of information out of these early tea leaves. Certainly, though, I wouldn't take them at face value.


  1. Jonathan,

    Why does Palin have to poll at 40% and not Romney or Huckabee? Their name recognition is almost as high as her name recognition. So you would expect them to be around where she's at right now. I'd agree that she would in trouble if she wasn't in the 30s against Tim Pawlenty and the minor candidates. But against candidates as well-known as she is in Huckabee and Romney? I don't see how you could expect her to be blowing them out by 20 at this time.

    Isn't the argument that what she gets now is all she's going to get completely spurious given that she's the top second-choice in many of the polls released? Assuming that what is reflected in these polls is that she's the strong second choice of Huckabee supporters, what's the problem for her?

    As for your straw poll point, the problem that you run into with your all of your analyses is that you think regular voters are often swayed by activists and that you assume that activists are the most hard-right. The activists that attend an event like CPAC are likely more socially liberal than regular Republicans as a whole.

  2. Palin is just much, much better known, which means that she's probably -- not for sure, but probably -- closer to her maximum potential than the others are.

    For example, the latest NBC poll had Palin with 2% don't know/not sure, and Romney at 22%. FWIW, Romney also had far more "neutrals" in that poll, 30%, than Palin (20%).

    I do think that just-plain-voters are swayed by activists (and pols, and opinion leaders, etc.). I don't have any assumption that activists are the most hard-right, but there is plenty of research showing activists tend to be more ideologically extreme than most voters.

  3. Jonathan,

    The NBC poll is pretty much a giant outlier compared to other polls. Around 80% of GOP and gop-leaning indies know Romney and Huckabee.

    See the CNN poll yesterday.

  4. Anon,

    There's hundreds of times more info out there about Palin than about Romney or Huck; she's off the scale compared to them. We're not talking about activists, here. Out there in the electorate, she's obviously much, much, better known -- not just more people have heard of her, but they know something about her, and feel they have enough info to make a decision. We're not talking about Romney/Huck levels, where people might vaguely recognize their names and will be willing to give an answer if pushed.

  5. but that's not reflected in the polls. Romney and Huckabee's name rec is almost as high among regular Republicans and republican-leaning indies.

    You may anecdotally believe they aren't as well-known based on google hits and media attention but all three are well-known with republican and republican-leaning indies. It's more than just vague recognition.

  6. Anon,
    I think the point is that opinions of Palin are more fixed than they are for Romney or Huckabee. You've got to get to the upper levels of political information to find people that know more than the basics about them. This is hard to get in a poll, because you'd have to use open-ended questions, and few pollsters do that.

    Yes, people recognize their names almost equally. But I suspect that many Americans could tell you a LOT more about Palin than either Huck or Romney. I tend to doubt it is much beyond vague senses with Romney and Huck. Romney.....isn't he the Mormon one? Huck....the guy who used to be fat, right? I just can't see someone asking a question when I ask them: Do you know Sarah Palin?"

    You can't make the case that 54.8 billion articles written about Palin haven't made a difference. (number made up) Coverage provides information. Yes, there's information saturation, but it can't be at the low levels of exposure Huck and Romney have had over the last 3 years.

    However, to NOT defend Jonathan, I'm not sure that all straw polls are created equal. I'd think that CPAC is likely a more representative straw poll than the Ames BBQ. Anecdotally, I have the impression that few people are "sent" to CPAC to stuff the ballot box, whereas that's practically the express purpose of the Ames BBQ.


Note: Only a member of this blog may post a comment.

Who links to my website?