The Rationalizing Voter

Political science paper discussion time! Woo! I was saving string on the Bartels/Achen paper (pdf) on how voters rationalize their decisions, but I decided to do my column on S-CHIP instead, so lets talk through things a bit here.

There are basically three theories of how voters think about issues when casting their ballot for a candidate. The first is "policy-oriented voting," in which a voter examines the various candidates on offer and votes for the one whose policy proposals most closely match the voter's own preferences. The second is "persuasion," in which the voter alters their own opinions to more closely conform with those of the candidate or party they favor. And the last is "projection," in which voters convince themselves that the candidate or party they favor really does possess the same beliefs that they do.

Sadly, persuasion and projection are much more common than policy voting. Voters aren't stupid. They're just much more committed to political parties and candidates than they are to issues. And we're not talking about the faceless rubes in the hinterlands. The effects of partisan bias actually become stronger as a voter's level of political information rises. High information voters, then, are actually more likely to warp their opinions based on their political allegiances. An example of this comes with the trajectory of the budget deficit during Clinton's first term. Bartels and Achen write:

the views of Republicans and Democrats diverge as we move from the bottom to the middle of the distribution of political information; partisan inference seems to dominate throughout this range, since the widening gap owes at least as much to Republicans moving further from the objectively correct answer as to Democrats moving closer to it. The pull of objective reality only begins to become apparent among respondents near the top of the distribution of political information. Among the best-informed 10 or 20% of the public, even Republicans were slightly more likely to say that the deficit had decreased than that it had increased, and Democrats -- untroubled by any contradiction between the facts and their partisan expectations -- were very likely to recognize at least some decrease.

The self-deception of high information voters is not, I think, much of a victory. In a basic way, high information means that you take in a lot of information. When that happens, you begin to pick out guides to help you make sense of the torrent of data and facts This blog, say. Or David Brooks. Or George Will. Depending on the orientation of those guides, you'll see different facts emphasized and a general impression emerge. If that general impression is that "Bill Clinton is a bad president who doesn't know anything about the economy," your sense of the deficit will probably be that it increased, even if you've not actually been lied to on the subject. As a high information voter in other words, you know enough to seek out information sources who will validate your world view -- which can then deceive you.

Bartels and Achen thus conclude that "Most of the time, the voters are merely reaconfirming their partisan and group identities at the polls. They do not reason very much or very often. What they do is rationalize. Every election, they sound as though they were thinking, and they feel as if they were thinking, as do we all. The unwary scholarly devotee of democratic romanticism is thereby easily misled."

This is different than saying that voters are irrational. Voters are quite rational. Indeed, asked about their political opinions, they tend to be quite consistent. When queried, in isolation, about policy matters, they tend to give reasonably well-informed answers that demonstrate impressive internal coherency. But come election time, that rationality is overridden by tribalism. So Achen and Bartels propose something of a distinction. It's not enough to merely have rational voters. You need civic competence, too. "Competence," they write, "requires not only logical consistency and cognitive efficiency, but also some modicum of accuracy in perception and receptiveness to new and, perhaps, disconfirming evidence." And that is what the voters simply don't demonstrate -- and nor do their guides, the hyper-high information pundits.

If you need evidence of how tribalism destroys a high-information citizen's person's receptiveness to contrary data and disconfirming evidence, you need look no further than the current conservative frenzy over the statements of Private Scott Thomas Beauchamp. In this case, some very smart, very high-information pundits on the right have managed to convince themselves that a series of statements are false not because they have discovered any falsehoods, but because Beauchamp's tales from Iraq do not accord with their beliefs about the military. All the time spent fact-checking and arguing over the issue makes it look like the braying fact-checkers are thinking and studying -- and to a reader, all that fact-checking and arguing almost certainly looks credible -- but in fact all they are doing is rationalizing their preexisting opinions. They only look like they're thinking