It's widely known that the Democratic Party rests upon a voting base of minorities - those who don't completely fit within the American mainstream. After the Civil War, the states in the former Confederacy voted Democratic. When Catholics were discriminated against, they voted Democratic.
Today the most well-known Democratic minorities constitute blacks and Latinos. President Barack Obama's coalition rested firmly upon these votes - a problem in mid-term elections, when these voters tend not to turn out.
There are, however, a lot of very Democratic minorities out there other than just blacks and Hispanics.
A few days ago I notice a pattern and did a little research to confirm it. It turns out if you know a state's population density, how many people are in it per square mile, you can make a good guess as to whether that state will give its electors to Obama or McCain. In fact, at the time I wrote that, Obama led in the top 17 most dense states or districts and McCain led in 9 of the 11 least dense states (and the other 2 were close). The candidates split the middle 23 states.
The findings spurred a lot of discussion and many commenters wondered if the relationship would be stronger at finer granularity because some states (like Nevada) may not be very dense but it's only because huge parts of them are entirely empty while most people live within the cities. Matthew Yglesias linked to the piece and suggested looking at the county level.
Dave Schor took him up on it and ran the analysis. It turns out there is a correlation at the county level as well. But a related finding was more interesting. A county's relative density compared to the state as a whole matters more than it's absolute density. In other words, the most dense county in a non-dense state may vote for Obama at a higher rate than the second or third most dense county in a very dense state. More details here.
STATE BLOGS
