I want to make it clear at the outset that I fully expect Obama to win this presidential election, but harbor no hopes of a landslide victory. I still think it's more likely that the Peach State will be lightly up red on the night of November 4th. Nonetheless, I do believe that Obama has a much better shot at winning Georgia than do most observers. Details are below the fold.
A post by Ben Smith caught my eye earlier today. In it, Smith quoted an Atlanta Journal-Constitution article regarding early voting among African-Americans in Georgia:
A disproportionate number of Georgia’s 194,138 early voters are African-American, in what could be an encouraging sign for Sen. Barack Obama’s presidential campaign.
As of Wednesday, about 39 percent of those voters — 74,961 — are African-Americans, Georgia Secretary of State Karen Handel’s office said.
African-Americans make up 29 percent of registered voters in the state, according to Oct.1 figures. They cast 25 percent of the total votes cast in the presidential election four years ago.
This report made me consider what type of African-American turnout current polls are projecting for Georgia.
I took a look at the two most recent polls of the state that included internal data on the racial breakdown of respondents. I would have liked to look at more data, but the recent polls for Georgia that provide decent internals are scant. The two polls I considered were a Survey USA pollconducted between 9/28 and 9/29 and an ARG poll conducted between 9/18 to 9/21. The results from these three polls are as follows:
Survey USA
677 Likely Voters
All Voters White (67%) African-American (25%)
McCain 52 73 4
Obama 44 23 94
ARG
600 Likely voters
All Voters White (73%) African-American (27%)
McCain 57 76 5
Obama 39 19 93
Obviously, both of these polls are assuming turnout rates similar to those in 2004. In both, 4% of respondents either chose other candidates or were undecided. Survey USA also included Hispanics (5%) and Other (3%) in their racial demographic breakdowns, but did not give data on these groups' support for Obama and McCain.
On the basis of the Atlanta Journal-Constitution article, I decided to calculate how the overall McCain-Obama numbers would look with higher levels of African-American turnout. I did some quick calculations to see how things would change if African-American turnout were 30%, 35%, or 40%, assuming that within each ethnic group the support for McCain and Obama remained unchanged. The results are as follows:
30% African-American Turnout
Survey USA ARG
McCain 46.9 54.7
Obama 42.6 41.2
35% African-American Turnout
Survey USA ARG
McCain 43.8 51.2
Obama 46.3 44.9
40% African-American Turnout
Survey USA ARG
McCain 40.7 47.6
Obama 49.9 48.6
For the Survey USA results, I split the Hispanic and Other vote 50-50, which is probably rather generous for McCain, as most national polls have shown Obama winning a substantial majority of the Hispanic vote. Regardless, Obama clearly has a shot at winning Georgia if the African-American turnout resembles what we have seen so far in early voting. His numbers are obviously much worse in the ARG poll, but it should be kept in mind that Survey USA is generally considered a much stronger pollster than ARG. (For example, Nate Silver puts Survey USA's pollster-induced error at +1.02 and ARG's at +2.32.)
I have no sense as to how early voting turnout among demographic groups correlates with their turnout overall, but I suspect that early voting is tied to the "ground game" a campaign is running. The high level of African-American participation so far seems to be indicative of Obama running a strong campaign in Georgia, which could also manifest itself on election day. The Obama campaign itself has seemed to think that Georgia is in play, but there hasn't been all that much talk of Georgia turning blue given the rather strong performance of McCain in recent polls.
Of course, this analysis is based only on an examination of two polls that are a few days old. A lot can happen in the next month, and it's still unclear how the Barr and McKinney candidacies will play out in Georgia. If I were putting money on the outcome of the election of Georgia, I would probably bet on McCain. But luckily I'm not, so there won't be any bitterness on November 4th if Barack Obama becomes the first Democrat to pick up the state since 1992.