Tuesday, 30 December 2003

Under the watchful eyes

Brian J. Noggle adds more potential subversives for you to be aware of, in addition to those evil Almanac-toting folks.

Master, Commander, and $9.25 for popcorn and a Coke

Dad and I saw Master and Commander today, which was most enjoyable—even if I could have done without the ER/Black Hawk Down at sea bits. Swordfights? OK. Arms being amputated with meat cleavers? Merci, non.


Lots of people—sometimes, even me—have trouble remembering where their keys are. Venomous Kate apparently has problems keeping track of her vibrator. I guess Hawai'i is more exciting than the mainland after all.

On partisanship

Ken Waight’s Lying in Ponds, which combines semi-regular “weblog-style” entries with daily analysis of America’s leading newspaper pundits, is one of the more worthwhile diversions in the blogosphere, even if I don’t share Waight’s apparent antipathy toward partisanship.

Partisanship has numerous functions in democratic society. In the electorate, it creates a psychological attachment between voters and politicians by providing a convenient “brand label” for voters, and a shortcut for voters who don’t have the time or inclination to research the qualifications of down-ballot candidates—even though it’s not immediately clear what meaningful difference there would be between a Republican and Democratic sheriff.

At the elite level, much of the role of partisanship is about communicating a consistent message to the public—the key part of Lazarsfeld and Katz’s “two-step flow” of political information. In essence, the public uses cues from elite figures to help inform and decide their own positions on political issues. Strong partisans, like Paul Krugman and Ann Coulter, are part of this process—as are more conflicted pundits like Tom Friedman. If there’s a risk associated with pundits like Krugman and Coulter, it’s that they give an exaggerated version of their own party line that often borders on caricature. But I don’t believe they are quite as harmful as those like Waight, who are bothered by the most partisan pundits’ apparent singlemindedness, seem to think.

Explanation, not prediction

Dan Drezner, subbing for Andrew Sullivan, discusses problems with forecasting models and the media members who latch onto them. One notable oversight in forecasting: virtually all of the existing models predict the nationwide vote, rather than the outcomes of state elections to the electoral college—a particularly problematic consideration when dealing with close elections, like that in 2000. The ones that do make state-level predictions are rather dated.

More to the point, as Matt Yglesias points out, aggregate-level models are often inherently problematic. The problem that Yglesias calls “specification searching”—or what I’d call atheoretical modelling, with a healthy dose of stepwise regression to boot—is endemic to the whole class of forecasting models, because fundamentally they are inductive exercises, focused on finding the best combination of variables to predict the observed outcome. Most good social science (or science in general, for that matter), by contrast, is deductive: establish a truly explanatory theory, develop specific hypotheses, and operationalize and test them.

That isn’t to say, however, that unemployment doesn’t belong in the model at all; it may, for example, be the best available indicator of a theoretical construct like “voters’ perceptions of the national economy.” But as someone whose research interests are more centered on individual-level explanations of behavior, rather than attempting to explain aggregate outcomes, I sometimes wonder if aggregate-level models trade too much scientific value for their parsimony.

See also James Joyner, who points out that small sample sizes aren’t necessarily problematic when the universe is also small. However, in a small sample the good social scientist will be particularly attentive to the potential issue of outliers—atypical observations that can lead one to make conclusions that aren’t justified on the basis of the data as a whole.