Wednesday, September 19, 2012

"Fundamentals" Frustration

John Sides isn't quite as frustrated yet about the idea that the economic "fundamentals" of the 2012 cycle make a Mitt Romney victory so inevitable that anything else needs to some fantastic explanation as he is at misreporting of independent voters, but I think he's getting close to breaking out the all-caps.

(Short version again: the prediction models are basically predicting a close race; as I read them, it's with Barack Obama as a slight favorite, but it really depends on how you look at it -- some models predict a Republican win, some a Democratic victory).

I think the level of frustration that John (and some of the rest of us) feel at this has a pretty simple cause: almost everyone who is open to persuasion on this point gets it, and so we don't get any more easy victories. That is, those who are open to reading what political scientists have to say -- and it's an impressive number of reporters, pundits, and practitioners -- are reading the Monkey Cage, reading the other political science blogs, reading evidence-based sites such as Nate Silver's, and taking what they read seriously. Indeed, I've been consistently struck since I started doing this by how many people really do want to learn what political scientists have to say.

The problem is that what's left are those who for whatever reason aren't very interested. There are lots of reason for it...their own biases, in some cases; in others, perhaps incentives for believing, for example, that election results are mainly driven by gaffes and other day-to-day campaigning events (not that campaigns are irrelevant! No one is saying that!).

All of which makes for a lot of frustration. But I do think that the flip side of that frustration is that a lot of progress is being made.


  1. In general I think the models on this issue should be paid more attention to, and yet......this IS the most severe economic downturn since the Great Depression. Isn't there some validity in believing that 2012 is outside the data of the last 50 years?

    Just because this level of economic growth = incumbent victory over the years of the Great Moderation doesn't mean the equation holds during this Depression.

    I'd say we are at least partly outside the bounds of (recent) past experience, and political scientists should recognize that even as they continue to use historical data. But then again,I'm biased towards believing that Obama is a better-than-good candidate which accounts for his continued success among moderates/independents and consolidation of the liberal base.

    1. Sides and Vavreck note that many people are willing to accept the argument that Bush is responsible for the state of the economy. And apparently, they just find Obama more likeable.

      Note, too, JB's link to Floyd Norris's article in the New York Times in which he points out that job growth was actually worse during the Bush years (negative in Bush's first term, slightly positive but still less than Obama's in the second), albeit from a higher starting point. Norris includes all the Bush years, including 2008, but I've seen other arguments that even if you only take 2001-2007 and leave out 2008, job creation under Bush was the slowest in postwar history.

  2. There is a reason people don't listen to political scientists. The map is not the terority. And for the vast majority of Americans electoral politics is just reality TV.

    In terms of Gafffe, the models dont' work for a few reasons:

    1) Gaffee suck oxygen out of the room. You've wasted 3-4 days on this. You've only got, what, 50 left?

    2) Polls lie. After hearing about the Gaffe of the day, how do you respond to the latest poll what calls you.

    3) true undecided votes are basically idiots. They watch CNN for 24 hours and try to decide how to vote. You're crazy if you don't think that gaffee play has an effect.

  3. For some folks, hesitation arises from skepticism about the statistical rigor behind the models. In particular, it strikes me that poli sci models assume away homogeneity of variance, a critical component of statistical inference, and one that - in my amateur reading - seems to be left out of professional discussions.

    For example, today we had a link to John Sides saying that Romney's 47% gaffe won't matter because clinging to guns and religion didn't matter, indeed, "gaffes don't matter"(TM). Perhaps this is so. However, when one considers that not paying federal income tax is usually a result of temporary circumstances, perhaps as much as 2/3 of the population has reason to be personally offended by Romney's comments; since Obama was speaking about abandoned rural communities, his comment surely personally offended no more than 1/10 as many folks as Romney's. Beyond that, while Obama's comment was patronizing, it was ostensibly sympathetic; Romney's cast its targets as hapless losers.

    Does that difference 'matter'? From here in the cheap seats, the two incidents look very different, different enough that it is hard to imagine inferences from one to the other. Unless, of course, you are in the professional business of "gaffe analysis", in which case both are "gaffes", no differences there, carry on and all that.

    I don't mean to be too hard on the profession, I think this happens in every part of the academy, after all models provide publications and grants and prestige and obviously can be very good friends. I think amateurs might be less skeptical if models were more connected to the real world, in particular, the statistical world.

    1. Ah, but CSH, to run with the insight (or, at least, what I take away from my own personal rejection of "bump-and-wiggle-ology"): for the gaffe to matter, I think three things would have to be necessary.

      1) A person hears about the gaffe.
      2) A person accepts that the gaffe is bad.
      3) A person's opinion of Romney changes sufficiently to change their vote intention.

      I don't think there are many people who fit all three categories. #1 is heavily skewed towards people like us: junkies (though, admittedly, few of them have as much of an addiction as folks on this blog do!). By and large, then, #1 is strongly correlated with NOT #3--Dems (like me) feel schadenfreude, Republicans have already heard the drum beat on the 47% over the last year. So, we're already down to the folks who may hear about this who are persuadable...not a very large set of people. Now, they have to find this to be particularly bad (#2). Problem is: in a presidential year, they're already sick of "all the attacks," and a number of them are getting to the "pox on both your houses" point. Toss on Romney's damage control (today's video from 1998, which is a whole different thing), and now we've got "he said, he said."

      Honestly, I expect last week's Libya error to have more of an impact than this, in that there was a solid day of Republicans running away from Romney (so partisans weren't given a reason to support their guy on it) AND foreign policy almost always plays to the incumbents advantage (regardless of whether or not the incumbent actually deserves it to). If Libya didn't shake 45.4% from liking Mitt, I don't think 47% will. OTOH, I could see the case being made that 2012 is all about domestic, so 47% would be the bigger one. Still, I come back to the 3 things and expect a small effect.

    2. A brick is not a wall.

      Persuading someone to vote is very different than making sure they show up.

      (in terms of economic models, they are also a bit off. The reality is there are a lot more unemployed (and credit starved) people out there. On the other hand, the rich are doing very well.)

  4. @Matt Jarvis, thanks for the insightful reply, I defer to your expertise, but nevertheless push back a bit on # 2: rather than concluding that the "gaffe" is bad, it seems to me the voter would have to change their opinion about whether said candidate is best for their interests or the country's. By that standard, most gaffes are obviously irrelevant: though the dozen or so Appalachian mountain men willing to vote Obama in '08 were no doubt offended by his "cling to guns and religion" condescending tripe, its hard to reconcile how that would have affected their prior decision calculus that Obama was the best guy for them.

    Small aside: it does seem to me that the pros tend to get a bit too "black box" with voter intentions, describing voter intentions based on exogenous factors (e.g. the economy) and less so what the voter thinks is best for them/the country. Even the most hardened, irrational partisan surely believes that this 15th vote in a row for their party's candidate is because he/she is in the voter's or the country's interest - this specific time - and not because said partisan is stuck in a wretched rut.

    So Romney told a bunch of rich guys, when the cameras apparently weren't rolling, that 47% of us are hopeless moochers. 47% in 2011, but since those in scope of no-federal-tax largesse shift from year-to-year, some higher percentage of us would have been in scope in the past decade. Broaden your definition to include those the voter cares about or are responsible for, and perhaps as many as 90% of us have reason to feel personally threatened by Romney's off-the-cuff (so presumably honest) comments.

    One other digression: imho liberals make way too much out of the conservative's paradoxical support of social programs (e.g. SS) and desire for drastically reduced federal spending. From a personal interest standpoint, I want government to stop spending money and I want the safety net if me or a loved one needs it. This is all rather logical at an individual level, if paradoxical at a 30,000-foot level.

    So, in conclusion, it seems to me that gaffes don't matter precisely because they have no bearing on the decision calculus of the average voter (i.e. what he/she perceives to be her own interest).

    By that standard, Romney's latest gaffe is Hall-of-Fame quality; it will be interesting to see whether it torpedoes his candidacy (where others have been harmless) on the theory that here we have that one, rare, miraculous gaffe that specifically announces to persuadable voters:

    "I'm really not the guy for you".

    1. @CSH: I wouldn't say that really is pushback, per se. Rather, we just have a difference of opinion over whether folks who fit #1 and #3 are likely to fall prey to #2 or not. I phrased it as "gaffe is bad" but it could easily have been (and maybe should have been) phrased as "is perceived to send negative information about Romney." So, I would consider your point a friendly amendment, but I just think that expectations like that are already baked into the electorate at this point.

  5. Well, it may not be gaffes or any particular campaigning event that moves the needle, but the quality and character of the candidate certainly matters.

    My dog could have defeated W in 2004, as he was an extremely weak candidate, but nominating Kerry was like manna from heaven for him. Kerry was easily caricatured as a flipflopping northeast lefty hack, and I'm not certain it even needed to be caricatured... he might have filled up the screen all on his lonesome. Romney is similar, weirdly so.

    I'd say in both of these cases, the caliber and quality of the challenger has mattered.

    1. Actually, based on fundamentals, Kerry did better than expected. So the "campaign effects" of the Kerry campaign were a net positive.

    2. ...that's too subjective an assertion.

      Again, my dog could have defeated W in 2004, but Kerry was easy meat, and got whacked.

      And I think my dog could have defeated Obama this cycle, but once again, a weak candidate with many of the same flaws as Kerry is poised to lose to a weak incumbent.

      The caliber and quality of the challenger does matter.

    3. Scott --

      Yes! I'm going to win this fight, one reader at a time...

    4. ...that's exactly what Kerry said. ;-)


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

Who links to my website?