On Wednesday, Nate Silver – in a piece playfully entitled “Is Mississippi The New New Hampshire?” – presented an interesting analysis of Gallup’s recent data dump on statewide changes in President Obama’s job approval. Silver rightly notes that the president’s job approval – measured against his 2008 vote – has ticked up slightly in several (mostly Southern) states. Meanwhile, his position in the South relative to the rest of the nation appears to have improved substantially. (You can get the gist of Silver’s claims from this easy-to-read chart.)
From these observations, Silver goes on to opine:
But, while it is all but assured that Mr. Obama will again campaign vigorously in North Carolina, Virginia and Florida, what about the other states in the South? Might there be any new swing states?
The most intriguing possibility is probably Georgia, where Mr. Obama’s approval rating was 45.5 percent in 2010 — just slightly below the national average. Compare Georgia, for instance, to Missouri: Mr. Obama lost the former by 5.2 percentage points in 2008, and the latter by 0.1 point. But Mr. Obama’s campaign heavily targeted Missouri, which it did not do for Georgia; that may have been worth several points. And Mr. Obama’s approval rating is now higher in Georgia than it is in Missouri. Depending on the identity of the Republican nominee and his or her geographic strengths, Georgia might be the better target for Mr. Obama.
Beyond that, there are a number of Southern states — South Carolina, Texas and even Mississippi — that Mr. Obama could plausibly take if he wins re-election in a blowout. But those states would almost certainly not give him the critical 270th electoral vote in an election that is otherwise competitive. So even if Mr. Obama wins a few more votes in those states, some of them could be wasted.
So, Silver’s claim is that changes in President Obama’s absolute and relative job approval suggest that his future presidential campaign might want to start looking South, because under certain conditions it could conceivably win some of these Deep Southern states.
I disagree with Silver’s conclusions. His findings about changes in Obama’s job approval are better explained by the peculiar political dynamic of the Deep South, rather than an actual improvement in his political standing.
First of all, a significant methodological issue. Comparing the Gallup poll directly to actual election results is a problematic, apples-to-oranges comparison. Gallup is a poll of all adults; by definition, voters, a subset of the adult population, decide elections. This could make a big difference in many ways. For instance, we should be careful in evaluating states where whites are heavily “over-represented” in the electorate relative to the population at large and they vote heavily Republican. In these states, sampling the adult population should mitigate Obama’s decline in job approval. Looking over Silver’s chart, that might explain why the president declined more nationwide than in Alaska, Oklahoma, Texas, and Utah – where whites were heavily over-represented in the electorate and gave McCain at least 65 percent of the vote in 2008.
Next, let’s think a little bit more about how President Obama’s job approval can change. Some groups, like independents, are going to drive that change while others, like white Republicans and African Americans, are not. Alter the balance of a state’s political composition between these (and other) groups, and Obama’s job approval becomes more or less vulnerable to alterations in the national mood.
This could be a factor in the Deep South, which, as the following chart details, has more Republicans and African Americans, and fewer independents, than your average swing state or the nation writ large.
Notice the small number of independents in each of these states (except Georgia). That is a strong sign that there are relatively few persuadable voters here, as these electorates exhibit pronounced racial polarization. Thus, Obama has little ground to lose here; he is basically locked into some number between 40 percent and 48 percent.
I think this helps explain the pattern of Silver’s data array even beyond the Deep South. If you look at his chart carefully, you’ll note that nineteen out of twenty-two McCain states are on the bottom half. (That pattern should be pretty evident in this map of relative change that Silver provides.) Few of the GOP states outside the Deep South have large African American populations, but they do have relatively large Republican populations. What's more, independents in Red States generally broke Republican in 2008, despite the strong headwinds the party was facing in 2008. In other words, Obama does not have as far to fall in Republican territory outside the Deep South, either.
So, in conclusion, we might say that Silver has it exactly backwards: Mississippi is not the “New New Hampshire;” it is instead the antithesis of New Hampshire, comprised of a largely static electorate that made its mind up about President Obama some time ago. Ditto, to varying degrees, most of the rest of the McCain states.