Tuesday, 8 June 2004

Emprical political science makes the Times

It’s not every day that you see a citation of The American Voter in The New York Times, but thanks to Nick Troester and Will Baude I stumbled upon David Brooks’ Sunday column on partisanship and rationality.

First, to settle the discussion between Messrs. Baude and Troester: Brooks’ analysis is essentially correct, although the transitory attachment voters would have with political parties under pure rationality wouldn’t be “party identification” (an affective—or emotional—orientation) as we conceive of it in American politics. Under pure rationality, voters would select among the platforms of the parties and vote for the party with the most desirable platform at that given moment, subject to the institutional rules governing vote choice (i.e. whether we are using plurality elections, proportional representation, majority-runoff, Condorcet voting, the alternate vote, or what-have-you, and the district magnitude).

Voting, I’d argue, has both expressive and (to borrow Baude’s term) instrumental aspects. One votes to both participate in the selection process—the way Downs conceived of voting—and to express preferences about how the government should act in the future. Much ink has been spilled over this debate over the past four decades (“proximity” versus “directional” voting, the rationality of turnout, etc.) and I need not recount it all here. Suffice it to say: voters aren’t rational in the Downsian sense (Page and Shapiro notwithstanding), people (to the extent they are rational) seek to maximize their expected utility, and Troester (despite his minor fault of not being an Americanist) is right—an outcome I attribute to Troester receiving a Michigan education, versus Baude’s Chicago one.*

On to Brooks, who shows he’s a little out of his field in his discussion:

Party affiliation even shapes people’s perceptions of reality. In 1960, Angus Campbell and others published a classic text, “The American Voter,” in which they argued that partisanship serves as a filter. A partisan filters out facts that are inconsistent with the party’s approved worldview and exaggerates facts that confirm it.

That observation has been criticized by some political scientists, who see voters as reasonably rational. But many political scientists are coming back to Campbell’s conclusion: people’s perceptions are blatantly biased by partisanship.

I’ll grant that he’s working in newspaper space, but there are a couple of caveats:

  1. I think he ascribes too much prescience to The American Voter on the role of perceptual screens or partsian filters. Most contemporary scholars would agree its psychological underpinnings are weak to nonexistent.
  2. Political scientists aren’t “coming back” to their conclusions; with the exception of the aforementioned Page and Shapiro, the Michigan approach has been essentially the dominant paradigm in American political behavior since around 1980, and was certainly a leading contender since the mid-1960s.

Still, this is about the best explanation of contemporary thinking on American politics you’ll find in about 600 words, and it dovetails rather nicely with Ken Waight’s work at Lying in Ponds on elite political discourse.

* I can say this because my training is about 3/4 Michigan school, by way of Ohio State, with a healthy dollop of Rochester-style behaviorism thrown in just for entertainment value.

Tuesday, 31 August 2004

The fuzziness of public opinion

Laura McKenna of 11D links and discusses an interesting New Yorker piece by Louis Menand on political scientists’ research on public opinion. It’s good as far as it goes (focusing largely on Converse, Fiorina, and Popkin), but I think it would help to have incorporated more recent research like Zaller’s Nature and Origins of Mass Opinion and Alvarez and Brehm’s Hard Choices, Easy Answers, not to mention the whole “affective intelligence” approach, all of which take issue—in important, but differing, ways—with the Conversian public incompetence thesis.

I’d also argue that Converse’s more important and lasting contribution was “Attitudes and Nonattitudes,” (1970) rather than “The Nature of Belief Systems in Mass Publics,” since I think most contemporary political scientists who study public opinion would reject the concept of “constraint” as an indicator of political expertise or competence.