Boris Shor writes:
An article by EJ Dionne about our red-blue paper (“Rich state, poor state, red state, blue state: What’s the matter with Connecticut?”) appeared today in the Washington Post. Dionne had called me on Friday to talk about the paper, and we had a nice conversation — he’s a very sharp guy and really understood what we were trying to explain.
A couple of thoughts about the paper…
0. The article is nicely written, and gets our major points across clearly. Kudos to Dionne, and I wish we academics could write so crisply. The following are a bunch of quibbles that academics like myself can’t help themselves from generating:
1. The title of the column (“What Kind of Hater Are You?”) is strange, at least as it relates to our paper. The title of the syndicated column (“Class Warfare and Political Polarization”) was closer to the mark, but still off. I guess the article had a point beyond highlighting our paper (gasp!), but the new headline seems needlessly provocative. Does the Post title op-eds, or only news stories?
2. Dionne loved our title, as do I — great job Andy. He asked if I had any children (not yet!), because the the title of our paper sounded to him like something written by a parent. I think that hit it on the head. Looks like Zacky’s (Andy’s kid) first research contribution! :)
3. Dionne doesn’t mention (not that he needed to) that we show a rather important shift in recent American political history: for the first time, wealthy states _and_ counties are now voting Democratic (see figures 1 and 5 from the paper). That is new (since 2000), and something to think about.
4. I’m not sure the intellectual leap from seeing an income component to voting to “class war” is self-obvious. That’s one interpretation, but another could simply be that the weights individuals assign to various voting factors differ by region/state/county.
5. He quotes the line “In poor states, rich people are very different from poor people in their political preferences. But in rich states, they are not.” A brilliant one — so simple, yet so profound! I wonder who wrote it …
6. Again, the result of varying slopes of income may not necessarily mean that the country is even more “polarized” or “divided”. It could just mean that people vote differently across the United States.
7. People love Andrew’s point about the availability bias leading to journalists’ extrapolation of the “rich Democrat” meme to larger patterns. I admit, I was skeptical on including it in the paper, so there you go…
8. I’m not sure I understand this paragraph: “Gelman and his colleagues help us understand why Southern Democrats such as Bill Clinton and John Edwards may be more attuned to the power of populism than northern Democrats such as John Kerry — and, perhaps, Hillary Rodham Clinton. Their paper also helps explain why Southern Republicans such as President Bush pursue policies that are hugely beneficial to their wealthy base even as they try to diminish the political impact of class warfare by shifting the argument to other subjects: religion, values or national security.”
We show (figure 6, page 13) that, in the South (except for VA, WV, AR, TN, and FL), rich counties generally support Republican candidates, while that doesn’t seem to be so for other regions (figures 7-9). Of course, the sparsity of data at the county level mean our coefficient estimates are noisy, but these broad patterns seem to come through. Also, figure 13 on page 19 shows that southern states like MS, AR, LA, AL have a high estimated slope coefficient for income (incidently, so do NM, WV, ME, not solidly red states).
I’m not sure about the link to populism, or governing/campaign strategy. Is Dionne saying that politicians from the South imbibe the local political culture of a high effect of income on voting, and then go out and campaign and govern accordingly? Given the necessity of cobbling a highly diverse winning electoral college coalition, I’m not sure this is a very powerful strategy. If true, wouldn’t the corollary of the argument implie that Democrats like Kerry should have paid no mind to economic issues, since in Maine, rich and poor counties don’t differ in their presidential vote?
9. Is it right to dichotimize voters as living in “two different political universes”? That’s the common thing to do, and we fall prey to the temptation ourselves, but I think there’s more of a continuum of regions and individuals than there are dichotomies (see figure 13, for example). Maybe our primitive brains naturally tend to draw sharp distinctions and thinking in continuous terms is unnatural, but I think the data support the latter.
10. Dionne, in his phone conversation, mentioned that we might be getting some funky results for Southern states for 1968 due to the Wallace vote (since we define Republican vote as a proportion of the two-party results), which sounds right. See, for example, Alabama’s 1968 vote by counties. In many of the counties, the proportion of the electorate voting for the Humphrey or Nixon tickets is very low, so the Republican proportion of the two-party vote may not be telling us what we want. Of course, this wouldn’t change our results.
11. It’s nice that the column is the most emailed, and the most viewed, on the Post today!
12. I’ve tried to collect all the press/blog mentions of the paper.
13. Why would a Washington Post column appear earlier in syndicated form, than in the original paper? The column was published in the SF Chronicle, for example, yesterday.
14. Could it be true that the Post collects comments at an AOL address?! (firstname.lastname@example.org)
15. Why is it impossible to buy a Washington Post in Chicago? “We don’t carry out-of-town papers”, the local Barnes and Nobles tells me. Those coastal hotshots can keep their elite opinions to themselves!