David Runciman writes,
Followed day by day, the race for the Democratic nomination has been the most exciting election in living memory. But viewed in retrospect, it is clear that it has been quite predictable. All the twists and turns have been a function of the somewhat random sequencing of different state primaries, which taken individually have invariably conformed to type, with Obama winning where he was always likely to win (caucus states, among college-educated and black voters, in the cities), and Clinton winning where she was likely to win (big states with secret ballots, among less well-educated whites and Hispanics, in rural areas).
“Predictable in retrospect”?? This seems like a contradiction. I agree with Runciman that there are patterns in the election results, but I’d only call it “predictable” if you actually predict it ahead of time, which I certainly didn’t! He continues:
The salient fact about this campaign is that demography trumps everything: people have been voting in fixed patterns set by age, race, gender, income and educational level . . .
Is that really true? My impression is that there are big differences between states, after controlling for demographics. See the New York Times animation here.
Finally, Runciman writes,
One of the amazing things about the business of American politics is that its polling industry is so primitive. . . . The recent New York Times poll that gave Obama a 12 per cent lead was based on interviews with just 283 people. For a country the size of the United States, this is the equivalent to stopping a few people at random in the street, or throwing darts at a board.
Hmmm . . . I wonder why they don’t just throw darts at a board, then? This would save them lots of money! For n=283, sqrt(p(1-p)/n)=.03, so +/- 2 standard errors is +/- 6 percentage points in the vote share for one candidate, or +/- 12 percentage points in the vote differential between two candidates.
Is this as good as a “dartboard” (i.e., an estimate based on prior information)? Maybe. It depends how good the dartboard is. The best solution is probably a weighted average of the “dartboard” and the poll, maybe weighting the dartboard more than the poll if the sample size is that small. But the size of the U.S. is not the relevant issue here. (The relevant formula is “1/n – 1/N.”)
The WSJ and the LRB; Yoo and Runciman
David Runciman teaches the history of political thought at Cambridge University. I’m surprised he didn’t walk down the hall and show his article to one of his political science colleagues there–I assume there are some researchers in American politics and public opinion he could talk with. This reminds me of John Yoo’s Wall Street Journal article which I discussed earlier. There’s no way that these authors intend to make mistakes; apparently, they don’t try a lot of fact checking. This seems funny to me–if I were writing for a such a large audience, I’d be petrified of making a mistake–but I guess that journalistic writing is a different world, even when done by academics.
P.S. I sent David Runciman an email asking for claification. If anyone here knows him (perhaps some U.K. political scientists read this blog) and could get him to explain what he was trying to say, I’d appreciate it.