In “Obama’s Race,” Michael Tesler and David Sears mount a strong case: Far from ushering in a “post-racial period” in American politics, Obama’s election “was more polarized by racial attitudes than any other presidential election on record and, perhaps more significantly, that there were two sides to this racialization: resentful opposition to to and racially liberal support for Obama.”

Another way to say this is that far from marking the end of us-vs.-them elections associated with Richard Nixon’s infamous Southern strategy, the 2008 election was arguably the beginning of its inverse: an electoral campaign where race, because of the skin color of the Democratic nominee, was a central issue, but this time, the “racially progressive” coalition proved larger than the racially conservative coalition. Call it the Northern strategy.

What’s interesting, though, is that the racial polarization has continued into Obama’s presidency. The chart above this post comes from a more recent research paper (pdf) by Tesla and Sears that concludes:

Very little has changed since Barack Obama became president. More specifically, we show:(1) Obama’s early presidential job approval ratings were influenced considerably more by racial attitudes than was the case for previous presidents, (2) support for Obama from white racial liberals had much to do with those highly racialized presidential approval ratings, (3) the effect of racial resentment on evaluations of Obama remained remarkably stable from early 2008 to November 2009, (4) President Obama continued to be evaluated not just as an African American but as someone who was distinctly “other,” and (5) Obama-induced racialization spilled over into issues on which the White House took visible positions, such as health care.

Obama wasn’t on the ballot in November, and so this particular advantage wasn’t present for the Democrats. But he will be on the ballot in 2012. And as we’re seeing in the Pew poll, it looks likely that the racial polarization — and thus his racial advantage — will persist. I wonder whether this is something that political scientists are taking into account when building their election models?