After the Debates, Reading the Electoral Map
I watched last week’s final presidential debate with GMF colleagues in town from Germany, Romania, and 
Serbia, which was fun and fascinating. Without attributing anything to them, they seemed to me to respond most to Obama’s demeanor: he appeared calm at a time when we’re using the word “crisis” attached to finances, Russia, Afghanistan, and other issues. McCain appeared forceful, which was effective at times, but also somewhat agitated (or perhaps frustrated). They were fascinated by the way the lines CNN used to record the simultaneous approval of a group of undecided Ohio voters seemed to reverse along gender lines: when McCain spoke, approval from men rose and from women fell; when Obama spoke, the lines reversed. And at the end, they turned and asked, “so what next?”
Read the electoral map, was my answer. As we’ve witnessed in the last two presidential elections, the state by state electoral count is key, not national polls. What’s remarkable is that the electoral map shows Obama winning outright at this point with 283 electoral votes without any of the so-called battleground states (see Real Clear Politics for a good map with polling averages state by state over time). And the number of traditionally “red” states that are toss-ups is also surprising – North Dakota, Nevada, and North Carolina — as are the traditionally solid states that are only “leaning” McCain, such as Georgia and Wyoming. There are still two weeks before the election, a lot of time in politics, but it suggests the potential for dramatic change in the “red state-blue state” map we’ve come to use to understand the United States.
Could this be a “realignment” as Charles Madigan wrote in the Chicago Tribune? Are the polls accurate? The best guesses here turn on two issues: potential racism and turnout. The first is known in the United States as the “Bradley effect,” referring to an election for mayor of Los Angeles when polls showing an African-American candidate ahead before election day proved unreliable, presumably because people were unwilling to tell a pollster their true voting intentions, that they would not vote for an African-American candidate, because they feared it would make them look bad. Andrew Kohut raised this possibility earlier this year when interpreting Obama’s unexpected loss in the New Hampshire primary. The counterargument is that the Bradley effect could be washed out if the many new voters registered by the Obama campaign actually turn out to vote and vote for him, but neither of these things are predictable.
All this suggests uncertainty: polls within the traditional margin of error of 3% should be treated with caution. Polls may be all we’ve got to guess how people will vote, but a lot of things seem up in the air in many directions at this point.
Two weeks to go!