Nice column by Michael Barone, "Adam Smith, Political Pundit". One cool bit:

It evidently irritates many liberals to point out that their party gets heavy support from superaffluent "people of fashion" and does not run very well among "the common people." They like to think of themselves as tribunes for the ordinary person, ready to spend the government's money to help him bear the travails of life, and they are puzzled when these people do not respond with proper gratitude.

Comments

[Andrew Gelman]

I have no problem with Barone's comments about Adam Smith, but his electoral analysis is wrong. He's making the common mistake of "personifying" states and counties. See here for a more detailed comments about his article on the "trust-funder left": http://tinyurl.com/3mqgk

Barone's an interesting guy but I'm disappointed on this one. I'd expect the editor of the Almanac of American Politics to be more informed about voting patterns. Perhaps the problem is that, because the Almanac is geographically-based, he's more easily misled into thinking that geography explains all. Richer voters support the Republicans but richer states support the Democrats. It's not a simple pattern.

[Craig Newmark]

Professor Gelman, thank you very much for your comment and for your links and further discussion on your blog. But with all due respect, I don’t understand your analysis.

1. You write that Barone is guilty in his “Adam Smith” piece of “personifying” states and counties. The only passage I can find that might qualify is the following:

"You can also see what Smith describe in today's politics. It's most visible in Wyoming, our smallest state. Wyoming is full of what Smith calls "the common people," and in 2004 it voted overwhelmingly for George W. Bush. He carried every county but one. That was Teton County, which includes Jackson Hole, the ultraexpensive resort inhabited primarily by what Smith calls "the people of fashion." This is what I called the trust-funder left in a column last March, and you can see their influence in the huge majorities for John Kerry in such trust-funder havens as San Francisco and Berkeley, Aspen and Telluride, Martha's Vineyard and Manhattan."

You seem to interpret Barone as claiming, “Jackson Hole voted Democratic, therefore the trustfunders who live there voted Democratic.” But I think that’s not at all what Barone is claiming. He is claiming “trustfunders” vote heavily Democratic and that trustfunders are a significant part of the electorate in Jackson Hole (and Aspen and Martha’s Vineyard, etc.), and that’s why, therefore, these places voted Democratic. I see no ecological fallacy here.

2. In your other discussion, you seem to place a lot of weight on the exit poll finding that in 2004 Bush won 62% of the vote among voters with incomes above $200,000. Shouldn’t we treat that finding with a lot of caution? For two reasons. One, might not some voters in that income bracket be reluctant to tell strangers what they earn? Who doesn’t report and what bias does that create in the result? Two, might not voters in that income bracket have, on average, higher opportunity costs of time and so vote in ways not subject to exit polling? In particular, they might be more likely to vote absentee ballots, and they might be more likely to take advantage of “advance voting” (which apparently showed a sizeable increase in 2004).

3. But since I don’t have any idea about either the size or the sign of these potential biases, I’ll ignore them. Let’s suppose that we are completely certain that in the population of high-income voters only 38% vote Democratic. Obviously, that means in a random sample of voters from that population we are unlikely to observe a majority voting Democratic. But equally obviously that’s not what Barone is claiming. He is discussing a most definitely non-random and probably very small subset of that population. I see it as no less valid as asserting, “Senior liberals arts and social science faculty at elite colleges and universities tend to vote overwhelmingly Democratic”. Or “leading members of the plaintiffs’ bar”. Or “star actors and actresses”. He is referring to people whose income derives largely—maybe exclusively?—from inherited wealth, and even more restrictively, whose inheritance is in, loosely speaking, the form of “trusts”. What fraction of high-income voters do these trustfunders constitute? Do they actually vote overwhelmingly Democratic? I don’t know; I’d be happy for you to provide some citations. But if they constitute, say, 10% of high-income voters, they could well vote 8 to 1, or more, Democratic—as some Upper East Side and Jackson Hole precincts did, I’d guess—while that population as a whole only votes 38% Democratic.

4. Note that if trustfunders do vote 80 or 90% Democratic, they could well be, as Barone asserts, puzzled and disappointed by the voting of lower income groups, who even while voting majority Democratic, vote a noticeably lower percentage Democratic than they do. I believe Barone is referring to the attitude captured in the famous comment of Pauline Kael: "I can't believe Nixon won. I don't know anybody who voted for him." (http://www.answers.com/topic/pauline-kael)

5. Finally, I realize that part of our charge as academics is to question authority. That said, though, I think when he speaks about U.S. voting patterns, Michael Barone should be granted the benefit of the doubt. Here’s a man who claims it’s not really hard to know a lot about U.S. voting; after all, there are “only 3,141-plus counties in the U.S. So you only have to keep a finite amount of information in your head.” (http://www.pajamasmedia.com/index.php?p=2005/09/barone_blog_michael_barone_edi.php)

[Andrew Gelman]

Craig,

Thanks for the comment on the comment. It's good to have an opportunity to clarify my thinking on this. In response to your comments:

1a. Here are two quotes from Barone--one from the first page of his article and one from the second page--that, I believe, personifies states and counties:

"Wyoming is full of what Smith calls "the common people," and in 2004 it voted overwhelmingly for George W. Bush."

"It evidently irritates many liberals to point out that their party gets heavy support from superaffluent "people of fashion" and does not run very well among "the common people.""

Here, I honestly don't think Barone realizes that the Democrats do better among the "commoners" (at least, as defined by income) and not so well among the more financially successful.

1b. I'm not addressing, one way or another, Barone's explanations of Jackson Hole, San Francisco, Aspen, etc. However, both at the national level and within states, the evidence is that the richer voters support the Republicans and the poorer voters support the Democrats. So, even if the Democrats are a majority in Denver, and Denver has some rich people, it's a huge leap--and a leap that's contradicted by the data--to say that the Democrats don't run very well among "the common people".

1c. Regarding Wyoming: yes, the richer counties in Wyoming are more likely to support the Democrats. This is also true in California, Oregon, Washington, and Idaho, among other states. On the other hand, in the south, the richer counties tend to support the Republicans. (see page 14 of this presentation:
http://www.stat.columbia.edu/~gelman/research/presentations/redblue.present.pdf
) A lot depends on where you look. With Barone's detailed knowledge of U.S. politics, I'm sure he could have easily picked affluent suburbs of Dallas, Minneapolis, Atlanta, etc., to find rich people who support the Republicans.

2. You're right that exit polls can have problems. However, the positive correlation between income and Republican vote is confirmed by every other poll I've seen, including decades of the very high-quality National Election Studies. It's real.

3. I agree that it is possible that the trustfunders, or super-rich, or whatever, are different than the other rich people. This comment was raised in my earlier blog entry, in fact. Here's my response:

"Yes, I agree that it's possible--I just don't think it's likely. Given that 62% of the rich folks voted for Bush, the trustfunders would have to be a lot different from the others for Barone's claim (that they are "very liberal politically") to hold water. I don't see the evidence to support the claim.

It's possible that Barone has this evidence, but it appeard from his column that he was making the claim based on data such as the Democrats getting 77% of the vote in Boston, 66% in Boulder, etc. So I think it probably is an example of the ecological fallacy."

You (Craig) write that Barone's statement is "no less valid as asserting, "Senior liberal arts and social science faculty at elite colleges and universities tend to vote overwhelmingly Democratic"." My reply is that Barone's statement is (potentially) much less valid in that he has presented no evidence! People have actually collected data on the political views of college faculty, business executives, college students, soldiers, and other groups. I don't know of any such study of trustfunders. The aggregate voting patterns at Jackson Hole, Aspen, etc., are not answering this question.

4. The Pauline Kael quote is funny, and the connection between voting and social networks is interesting. (My colleagues and I are actually in a midst of a study on the topic.) I think that Barone, and also the modern-day successors of Kael, all would benefit by coming to grips with the fact that 50% of the population of the U.S. disagrees with them! To quote myself again: perhaps the Democrats are the party of trustfunders, welfare cheats, drug addicts, communists, and whatever other categories of people you don't like. Perhaps the Republicans are the party of rich CEO's, bigots, fascists, and so forth. No matter how you slice it, both sides have to add up to 50%, so you either have to throw in a lot of "normal" voters on both sides or else you have to marginalize large chunks of the population.

I think that partisans on both sides want to believe that the people on the other side don't deserve our respect. And I don't like to see that from Kael or from Barone.

5. Yes, Barone knows a lot--which is one reason why I was moved to react to his column. Even the best of us can make mistakes, and I conjecture that Barone's very focus on localities makes him more likely to succumb to the ecological fallacy.

The thing that's frustrating to me about Barone's column is that . . . there really is something interesting going on out there (which is why my colleagues and I are researching the topic). All the evidence finds that richer voters support the Republicans. But richer states support the Democrats more. So counties, and states, really are meaningful. The political culture in different places is different, and voting is not determined by income. By picking out a few rich areas that support the Democrats, and then thinking that the "commoners" support the Republicans, Barone is missing a lot of intersting stuff out there.

[Kyle N]

Why do you think the Lefties spend so much time on the Enviroment and Animal Rights? Because these are constituencies which cannot say to them, "We don't need your help, go away".