It turns out that one variable that does a good job explaining how well Obama did in a state is how many people identify as white evangelical or born-again christian. Below, we see that Obama's support in a state is closely related to the percentage of voters identifying as white evangelical; on the right, a map of Obama's support among white evangelicals (purple shows the lowest support for Obama):
Click to enlarge.
Below, more details, more maps, more relationships, and how some of this fiddling with numbers might show that the reason Obama underperformed in certain areas is because he didn't campaign there.
The White Vote
Charles Franklin recently showed us how Obama did among white voters as a function of the percent African-American in the exit polls, and compared this to how Kerry did. Here's that graph reproduced:
Click to enlarge.
To no one's surprise, both Democrats did terribly among whites in the Deep South (Kerry a little better than Obama), which has the highest percentage of African-Americans, and there was a decent relationship between the two variables in states with more than 20% African-Americans (see the right side of the graph). But in other states, there is a huge variation in the support for Obama, with no clear pattern on this graph. What is up with those crazy white people?
Moving on to White Evangelicals
Let's go back to the graph from the introduction, and add a map showing what percent of voters in each state were white evangelical:
Click to enlarge.
All data are from exit polls or election results. The dashed red lines show values from the entire country (43% of whites voted for Obama, and 26% of the electorate was white evangelical). Also, note that I'll be using 'white evangelical' as shorthand for 'white evangelical/born-again christian' (although the acronym, WEBAC, is kinda cool).
Most states fall pretty close to the blue regression line. DC is pretty high above the line. States where Obama performed worse than we might have expected based on the population of white evangelicals alone are Alaska and Arizona (home states for the GOP ticket) and Idaho and Wyoming (significant chunks of Mormons in the population). Utah would probably be an outlier on there too, but there's no data in the exit poll for white evangelicals in Utah. The states with the most white evangelicals are AR, WV, TN and OK, all with more than 50%. The states with the least are MA, HI, and NY. Remember, some states have no data.
Back to All Whites
Here's a map of how whites voted, and, in the middle, how the entire electorate voted using the same color scheme for comparison. On the right, the percent of whites in the exit polls by state.
Click to enlarge.
The base of Republican support is pretty obvious: whites in the Deep South, and the triumvirate of UT, ID, and WY. This map helps give a little perspective to the red tone on total vote maps, especially the county-by-county ones such as in my last diary. Whites in West Virginia, for instance, despite the state having trended quite red over the last 20 years, still gave Obama about the same percent of the vote as whites nationwide. Whites in Mississippi are about as solidly Republican as you can get, on the other hand, despite the state as a whole giving Obama about the same support as West Virginia.
Evangelical Whites as a Portion of the White Electorate
So the maps above show us that when comparing states we need to look at demographic composition. So let's redo the graph above, but this time as the y-axis we'll use Obama's support among white voters only, and we'll use on the x-axis the percent of white voters who are evangelical, not the percent of all voters. On the right it is a map showing what percent of whites are evangelical by state.
Click to enlarge.
We see a huge range here, from three quarters of all whites identifying as evangelical to almost none. Not surprisingly, we get a better regression than before: there is a very strong relationship between the proportion of whites who are evangelical and Obama's share of the vote among whites. We still have states with significant Mormon populations (WY, ID) and home states (AZ, AK) below the regresssion. We also see that LA is now pretty far below the regression, as the only Southern state with a substantial non-Hispanic Catholic population.
Down the Rabbit Hole
Let's get a little more complicated. Now that we know how white evangelicals voted, we can compare it with how whites who are not evangelical voted. Here's the two maps for the geographic comparison:
Click to enlarge.
There's clearly a huge difference, yet geographic trends are roughly the same. Indeed, it's interesting that Obama was able to get support up into the 30s among white evangelicals in some states. This could be partly because of geographic distributions of different evangelical denominations (or even denominations with 'Evangelical' in their name that are not by most definitions actually evangelical), some of which may be more conservative than others.
Here's a graph showing the same data:
Click to enlarge.
The dotted line shows a 1:1 relationship. Again, support among white evangelicals is much lower than among whites who are not evangelical. But there is also an excellent relationship between the two. And still, we see whites in the South - evangelical and non-evangelical - show the weakest support for Obama.
Above, we saw the South has the highest proportion of whites that are evangelical whites. Let's see how support for Obama varies with the proportion of whites who are evangelical:
Click to enlarge.
As the proportion of whites who are evangelical increases, support for Obama decreases among both evangelical whites and non-evangelical whites, although it decreases faster for non-evangelical whites. Which leads to looking at the difference:
Click to enlarge.
White evangelical and non-evangelical support for Obama tends to draw closer together (although the correlation is weak) as the proportion of whites who are evangelical increases. However, the ratio of non-evangelical to evangelical support is constant at about 2:1.
Anyone still with me?
Summary
As the proportion of voters who are white evangelical increases, the support for Obama falls. Not a surprise.
As the proportion of white voters who are evangelical increases, the support for Obama decreases among all whites, white evangelicals, and white non-evangelicals. The gap between white evangelicals and non-evangelicals also decreases.
The question is, why? Here's some speculation. Perhaps because the white evangelical culture influences the media and social opportunities available to non-evangelical whites in areas where it is prominent. And, as a corollary, where white evangelicals are a minority among whites, they are influenced by non-evangelical whites. If this is the case, it is not an even-handed phenomenon. If it were, we would expect white evangelicals in places where they are a distinct minority to be much closer in voting behavior to non-evangelicals. That is, we would expect the last graph above to have a peak, falling to small numbers at both small and large values. This is clearly not the case, perhaps because where white evangelicals are distinct minorities they are nonetheless able to function within a white evangelical social sphere centered around worship and limit contact with non-evangelicals. I've seen this happen anecdotally, but have no numbers to match anecdotes with data.
They Are Not Lost?
In the same post linked at the beginning of this diary, Charles Franklin has a graph that shows the only states that had more than a point or two decrease in support for the Democratic presidential candidate between 2004 and 2008 were MS, AL, LA, and AR. We know that in three of these states, evangelicals make up the vast majority of white voters, and in LA they make up 49%.
So let's look at how the evangelical vote has changed since 2004. Overall, the share of the total electorate increased from 23% to 26% (enthusiasm for Palin?), but whereas Kerry only won 21% of them, Obama won 24% (so much for the Palin effect, I guess). The gains were not uniform, however, and in the South generally Obama underperformed Kerry among white evangelicals. Let's look at the map:
Click to enlarge.
Don't worry about Iowa: the difference is actually zero. Looking at the darker green or purple states, we see that most but not all of the darker green states are states where Obama campaigned heavily: NC, CO, IN, and OR (primary). In ND and NE, Obama campaigned (but did not visit much) and McCain did not.
However, of the darker purple states, Obama has visited AR, LA, and MS 5 times or fewer, according to the Washington Post candidate tracker. Georgia logged 9 visits, and a campaign presence.
So we have Obama gaining in NC (and possibly gaining a little in SC) but losing support in MS, GA, LA and AR among Southern white evangelicals compared to Kerry. One possible explanation is differences between white evangelicals in the Carolinas and in the rest of the South. Another possibility is that where Obama campaigned, he managed to increase his support among white evangelicals compared to Kerry. This last idea would add to the evidence in my previous diary, which also was consitent with campaigning leading to improved results. And that leaves me encouraged that a Democratic presidential candidate could make some gains even among Mississippi's white evangelical voters.
Cross posted on Open Left.
Diaries in this series (updated list):
Looking Back
Alternate History
Why Republicans Should Be Really Scared
African-Americans – We Are Not All of Us Alike
East and South Asian Americans – Diverse and Growing
West Asian Americans – Rapid Change
Native Americans – Increasing Participation
Islander Americans – In Need of More Representation
Native Alaskans – An Economic Factor?
Latino Electorate – Increasing Influence
European-Americans – Tribal Politics Persist
“Americans” – You Might Be Surprised
Appalachia – Surprisingly Democratic
Why Republicans Should Be Really, Really Scared
Why Republicans Should Be Really, Really, Really Scared
A Few More Tidbits
Related 2008 electorate diaries:
Maps: Blue America and the Changing Electorate
Maps: Obama and White Evangelicals