If you want to understand without a great deal of effort why the 2012 presidential contest is so close and how it might turn out, you should read two articles right away: Ron Browstein’s latest column for National Journal, and a piece published at TNR by Nate Cohn earlier this week. They both say basically the same thing: Obama is retaining much of his 2008 levels of support from minorities and college-educated whites, while slipping significantly from his already poor levels of support from non-college educated whites. Fortunately for the incumbent, the “new coalition” voters are making up a steadily larger part of the electorate, while the blue-collar whites (particularly men) who were once synonymous with the Democratic Party are becoming a steadily smaller part of the electorate.
This dynamic goes a long way towards explaining why Obama is competitive in North Carolina but is surprisingly vulnerable in places like Wisconsin, Minnesota and Iowa: his support was already about as low as it can get among non-college-educated whites in NC, but they represented an important part of his 2008 vote in the midwest. Poll variations largely reflect different assumptions about turnout among different categories of voters, with Republicans still hoping an economy-driven “discouragement factor” will depress turnout among Obama’s “new coalition” voters, and Democrats still hoping conservative lack of excitement about Mitt Romney will affect GOP turnout.
Moreover, this way of looking at the electorate is driving a lot of strategic decisions by the campaigns. The importance of placing a floor under losses among non-college-educated white voters helps explain the Obama campaign’s (or more accurately, Priorities USA’s) focus on Romney’s wealth and taxes. You get the sense the Romney campaign buys the idea (shared by many political scientists) that non-college educated whites will make up a disproportionate share of the undecided vote, and will overwhelmingly break towards the challenger late in the campaign if the economy does not significantly improve.
What is harder to assess, via either polls or public “messaging,” is how good a job the two campaigns and their parties are doing in voter targeting and GOTV preparations, though reportedly the Obama campaign sank enormous resources into these tasks early on.
In addition to the pieces by Brownstein and Cohn, you might also want to take a look at Alan Abramowitz’s first presentation of his 2012 election model, which–surprise, surprise–suggests a very close election, with one important variable–second quarter real GDP growth–unavailable until later this month.