Our Patchwork Nation

by John Sides on November 21, 2010

in Political science

The journalist Dante Chinni and political scientist James Gimpel have published a new book, Our Patchwork Nation, based on the Patchwork Nation project that I blogged about previously. Here is the book at Amazon. Here is a Washington Post review.

I asked Jim some questions about the genesis and findings of the book as well as the value of studying political geography.

Q: How did the Patchwork Nation project get started?

A: Dante Chinni, my coauthor, and then a reporter at the Christian Science Monitor and now at the NewsHour, wandered into my office in the Spring of 07. He was looking for a social scientist with an interest in the nation’s political geography. He had skimmed my earlier book, Patchwork Nation, and he had been bothered for some time about the journalistic reliance on the red-blue scheme for characterizing the politics of places.

As a professional journalist, Dante had a several main goals:

  • To get campaign coverage for the 2008 election out of the newsrooms in Washington, New York and Boston, and get it “out there” to the places where voters would be experiencing the campaign.
  • To build a longer-term online presence that would encourage other news organizations to move beyond their newsrooms as they cover the nation’s politics. We are now at Patchworknation.org, and we have a regular presence on the NewsHour’s site.
  • To eventually write a book, mainly a work of journalism, summarizing the 2008 election and its aftermath from the standpoint of representative communities that brought distinctive viewpoints to bear on the nation’s politics.

These goals required that we figure out what “places” should be covered, or represented. After all, no one could afford to go everywhere, and it probably wasn’t necessary to go everywhere.

But to which places? Somehow we had to segment the nation in socioeconomic terms that would offer more detail than red and blue but not be hopelessly complex for news consumers, the way many market research segmentation schemes are.

After some deliberation, we decided to try to characterize and classify locations drawing on county level data. Counties offered a number of advantages that outweighed their disadvantages. They were a convenient and low cost data source, widely understood and recognized by local populations, their boundaries rarely changed, and a wide variety of social and economic data were available.

After a couple of months of data exploration using pretty standard statistical tools, we settled on 12 segments, or “types of places” that are set forth in the book. Of course reducing each county to just one main category was a really difficult thing to do. We all know that many counties are internally heterogeneous and in fact we have a score for each of the 12 factors for each of the 3,141 counties. But for the sake of making the project intelligible to a mass audience, we had to try to assign a ‘best fitting’ category. This is where social science analysis had to compromise with the needs of journalism, jettisoning detail and precision for the sake of simplicity and ease of communication. Naturally, each of the resulting 12 categories has its own distribution, and readers with social science backgrounds will quickly realize that some counties fit their assigned type better than others. And it’s certainly possible that I have misclassified some. But I think on the whole we probably have it right. We do report on some validation efforts in the appendix to the book.

I should also add that the project has been greatly assisted by sizable and generous grants from The Knight Foundation.

Q: What do you think are some common misperceptions about American political geography?

A: Lots of people don’t think place matters, or that there really is no meaningful local variation in the way people learn or think about politics. And it is true that many aspects of place-to-place variability are not very important to politics. They key for social science is to identify which ones are.

I think the economy chapter in the book is a good one for that. Pretty much everyone agrees that “the economy” was the overriding issue in 2008. If we were to ask the “most important issue” question of 1,200 voters, it’s sure to be at or very near the top this year too. But people’s economic concerns may be remarkably variable because economic conditions are so geographically uneven. For some people, it is the foreclosure crisis. For others it is unemployment. For others it might be the uncertainty associated with retirement accounts or pensions.

We were interested to see that the most rural areas of the country have experienced neither high unemployment, nor high foreclosures. Residents there may be concerned, but they do not look around and see foreclosure signs, and unemployed neighbors, the way folks in other parts of the country do.

Another example: this year, there seems to be a remarkable spatial correspondence between the geography of the foreclosure crisis, and the geography of the Tea Party movement’s strength and activism. I don’t think I’ve pinpointed the causal connection quite yet, but I don’t think it is purely a coincidence, either.

Q: What do you see as the central findings in the book?

A: As a work of journalism, primarily, informed by social science research, the mass audience will “find” different things to be central than social science experts. For instance, I think a lot of the discussion in the culture chapter will come as a surprise to the mass audience, but will already be well-known among social scientists. That’s to be expected. To paraphrase the President, let’s be clear…this book is not primarily directed at academic social scientists.

I think that where I learned the most, or what I was most surprised by, is that people in different places have very different attitudes about saving, debt, and borrowing. The place where you live socializes you not only politically, but also economically. This is one reason why rural areas and smaller towns are not in the terrible economic fix that we see elsewhere. These populations are very conservative about matters such as taking on debt and borrowing. Moreover, so are their local institutions. Small town banks didn’t make risky loans on crazy terms. Our “Tractor Country” locales may not have experienced the boom that other places did, but their attitudes toward debt have certainly saved them a lot of grief in this ugly and prolonged downturn.

Q: What is important about the social scientific study of political geography?

Geography helps us not to forget the error term. We are very fixated on central tendency – what the regression coefficient tells us, but we are usually less interested in the standard error – exhibiting the extent of variation around that central tendency.

Making maximally general statements about politics and discovering law-like generalizations is important. But some of those generalizations are a lot flimsier than others. It is only through the close and continued inspection of variation, both spatial and temporal, that we come by an appreciation of how soundly we are generalizing.

Q:Where do you see this field of study going in the next several years?

A: I think a medium-sized wave is building in political science and other social sciences. When I started, I couldn’t get anything published except in really specialized academic journals. Within political science there were only a tiny handful of us. It’s easier now, not just because geographic approaches have gained wider acceptance, but also because we are better at framing our work, our statistical methods have improved, and we have persuaded more people of its importance. Having larger data sets available has been really key. It is easier than ever to investigate spatial variations in attitudes and behavior, using multi-level models, for example.

So I don’t see it subsiding. There will always be critics and skeptics, and I certainly don’t see it taking over political science. But I think we have gradually cleared a space for our work, and that it will become more established and integrated into the field as suitable data become ever more available.

Comments on this entry are closed.

Previous post:

Next post: