Before Bob Cushing and I published our first article in the Austin newspaper on sorting in 2004, we heard Dr. Fiorina wouldn’t like it. No kidding!!
Dr. Fiorina did not include figures from the 2008 election in his review (written in 2012), so the least I can do is to fill in that hole. The sorting trend continued in that election.
I know enough of how Bob constructed the analysis to get me in trouble, but I can tell you that “landslide counties” was not the way we measured this phenomenon internally. (Landslide counties was a device we created to illustrate the trend.) On the advice of several political scientists, we measured the change in the standard deviation of the Republican vote in presidential elections at the county level, weighted for population.
That weighted standard deviation of the R vote at the county level decreased every election but one (1964) from 1948 to 1976, for a total decline of just under 24 percent. It increased in every election from 1976 to 2008 but one (1988) for a total increase of 56.9 percent. The weighted standard deviation of the R vote increased from 2004 to 2008 by over 5 percent.
We didn’t use voter registration, that’s true. Voter registration shows a vast increase in political independents — something that political journalists say over and again, much to the consternation of writers here at The Monkey Cage. If voter registration is the way to go, then, indeed, there is a difference of facts.
But, really, there is a lot of interesting work in this area. We have been amazed at how many measurements move in the same direction.
James Gimpel began it all with Patchwork Nation. He, Iris Hui and Wendy Cho have new papers circulating showing sorting at lower levels than the county. See also this article by Seth McKee and Jeremy Teigen. Ian McDonald while at Duke and now in Portland State has done some very sophisticated and clever work using data from the U.S. Postal Service to measure the preferences of individual movers. He, too, finds sorting. (Also, Bob Cushing used IRS migration data to show the flows of people between counties. People from blue counties moved disproportionately to blue counties. Red to red.)
Ed Glaeser long ago saw the geographic clustering of people by education, yet another trend that began in the 1970s. (Roberto Gallardo at Mississippi State recently completed an analysis showing that this trend continued through 2010.) There have been recent reports (from, among others, the Institute for Health Metrics and Evaluation) showing that long-term trends that had every part of the U.S. enjoying increasing length of life have changed. In over 500 counties, longevity is now decreasing — again, a trend that began showing up in the 1980s. (There have been smaller studies showing a similar appearance of geographic differentiation in suicide rates.)
Robert Lang has shown the increasing relationship between population density and the presidential vote (something that was not present in the 1970s). Peter Francia wrote a very good paper [no link] on the turn of rural areas to Republican candidates (also a phenomenon that arrived since the 1970s).
Ron Lesthaeghe found that the relationship between family formation patterns at the county level and party preference increased from the 2004 to the 2008 election. Diana Mutz has written great stuff about the increasing political like-mindedness of churches (and the relatively mixed politics of the workplace). She also wrote in Hearing The Other Side about the more subtle kinds of segregation that is occurring in American communities.
David Schkade and Cass Sunstein have written about the psychological consequences of these slow-moving changes in residential living patterns.
Larry Bartels, of course, wrote the first paper (I’ve seen, anyway) on the increasing attachment of people to parties, a trend that also began in the 1970s. And Alan Abramowitz is just now publishing a book showing the increased polarization of Americans.
I’ve missed some, I know, but this is a start on what is a very interesting area of American life.