When D.C. voted 92.5-6.5 for Obama in 2008, it trended R

[Speed of Sound]

I know why this is the case (a simple maxing out of what a lopsided voting district can do in such a dramatic national swing), and I guess I don't have much to offer besides pointing it out, but it was something amusing I found while browsing some results and thought I'd share.

[Boston Bread]

San Francisco trended R more than 5 points in 2008 for the same reason.

[Nym90]

The "trend" calculation breaks down in extreme cases such as this. It simply calculates whether the nation as a whole swung toward one party more than the state did, but the more lopsided the totals are, the harder it is to continue to swing toward the majority party, as there are fewer votes to be gained.

A alternative way to calculate it to avoid this issue would be to determine the percentage of votes that swung of those who could possibly swing. In other words, a drop from 10 percent of the vote to 5 percent would be considered as much of a swing as a drop from 20 percent to 10 percent, instead of only half as much.

[Miles]

Yeah MA also trended R in 2008 because Obama didn't have much room to improve over Kerry there.

[nclib]

DC has always trended the opposite of the national swing.

[Xing]

Oklahoma trended D in 2012, even though Obama lost it by a greater than 2:1 margin. The most polarized parts of the country are usually guaranteed to "trend" the opposite way of the rest of the country, especially in years where there's a large swing toward one party or another.