Obama & Race-Based Voting: Exit Poll Data

I wanted to roughly analyze whether racism has been/will be an issue in the fall election. So I combed through CNN's exit polls, and pulled the following data for the question "Is race an important factor in your vote?" Note that this includes "Yes most important factor" and "One of several factors" together. Below is all the states for which the question was asked. I should also note that this obviously is data pulled from the Democratic primary. A general election universe of voters would obviously be more conservative.

The first column lists the state, the second the % of exit poll respondents who considered race important in their voting decision, and the third the % of voters who subsequently voted for Clinton (as in, this hurt Obama).

So, what did I find? It's easier to look at the states in groupings:

Race probably helped more than it hurt
Alabama / 29 / 35
Delaware / 18 / 28
Georgia / 21 / 24
Illinois / 23 / 28
Louisiana / 25 / 38
Mississippi / 31 / 36
North Carolina / 18 / 35

Race probably helped as much as it hurt
Connecticut / 15 / 41
Vermont / 13 / 42
Indiana / 16 / 47
Massachusetts / 16 / 51
Missouri / 19 / 46
New Jersey / 19 / 47
Tennessee / 21 / 52
Texas / 19 / 52
Wisconsin / 13 / 46

Race probably hurt more than it helped
Arizona / 14 / 56
Arkansas / 18 / 68
California / 17 / 61
New Mexico / 14 / 59
New York / 18 / 56
Ohio / 20 / 59
Oklahoma / 20 / 64
Pennsylvania / 19 / 59
Rhode Island / 18 / 65
West Virginia / 22 / 82 (!)

A couple of things to keep in mind:

1 - This is a cross-sample of states based only on CNN's exit polls, does not include caucus states since they do not do exit polling, and excludes Utah since only 9% considered race an important issue (no meaningful numbers could be extrapolated from that).
2 - Generally, it seems, race helps Obama in states with large African American populations. As with where it hurts, it seems that there's no easy answer.

So having seen these numbers, what do you think? Should only the percentage who said it was important matter, or should we also look at the percentage who said it was important and also voted for Clinton? Clearly, it's a non-insignificant issue; the average % who say they mattered was roughly 18-20%.  

Exit poll data taken from http://www.cnn.com/ELECTION/2008/.

Tags: data, exit poll, race-based voting (all tags)



In November, it'll be simple:

We won't win West Virginia. But we can do the following:

Motivate the African-American base to the point where in numerous states (Georgia, Louisiana, North Carolina, Virginia), only 1/3rd of whites will be needed to win the state.

It basically breaks down like this.

Places with lots of whites and few blacks and Obama is still favored positively; there aren't racial tensions.

Places with lots of blacks and few whites favor Obama; they feel he's a glass ceiling breaker, moreso than women feel Senator Clinton is. I'm not afraid to admit it is racially based but it's positive, not negative.

Places with an even number of blacks and whites, will mean a harder time for Senator Obama - whites feel there are racial tensions and it shows. Blacks vote for Senator Obama and whites for Senator Clinton. We'll have to see if whites continue to vote against Senator Obama in those places.

Then again, there are just demographics that favor Senator Clinton. More conservative areas, not conservative states, meaning outside the metropolitan areas.

More or less a reverse-Bradley effect or whatever you want to call it.

by Lord Hadrian 2008-05-16 11:33AM | 0 recs
Re: In November, it'll be simple:

Politico linked to a report today indicating that there are 500,000 unregistered, eligible AA voters in Georgia. And, Obama is doing everything he can to register those folks.

And as Poblana has shown, a significant increase in AA turnout can make a big difference. The same can be said about other parts of Obama's base, like younger voters. I don't think we'll win Georgia, but Virgian and NC seem plausible.

by DPW 2008-05-16 11:47AM | 0 recs
I did a breakdown pre PA

and found that if AA's make up more than 42% of the electorate the voters based on race broke in his favor, less than that and hillary got a boost from race based voting.

I would wager than in NC it was pretty close to even as for who it helped and it hurt Obama in IN and WV.

by Student Guy 2008-05-16 11:36AM | 0 recs
Re: I did a breakdown pre PA

In North Carolina it definitely helped him -- 64% of those who cited race as an important issue voted for him. In Indiana, you can't really say either way, and in WV it definitely hurt.

by VAAlex 2008-05-16 11:38AM | 0 recs
Thanks for the info

I didn't go and dig into the exit polls, I was guessing based on the composition of the electorate.

Based on the NC and IN results I think the math formula would go down to about 34% to the break even point (averaging all contests, LA really helps to bring this up).

by Student Guy 2008-05-16 11:45AM | 0 recs
Clinton Wins

Clinton wins every single state where the third number is 47 or more.  Obama wins every single state where the third number is 46 or less.

by soccerandpolitics 2008-05-16 11:54AM | 0 recs
Great diary

Finally, some helpful racial-issue dialogue!

I think we have to also account for the fact that, in the primaries, there are two factors at play:

* Democrats gave up being the racist party years ago with the Dixiecrats, so they may under-represent the racially-interested negative constituency.

...countered by...

* The crossover voters who (whether Limbaugh-inspired or not) voted in the Democratic primary because they wanted to make sure a black person did not become president.  

As an aside, Rush Limbaugh did run a musical segment "Barack the Magic Negro" sung to "Puff the Magic Dragon" several times, so they were, in fact, actively courting the racist demographic for Operation Chaos.

Regardless, when the general opens up, we're going to have to probably deal with a larger racist percentage (unless the second factor above totally outperforms the first).  One thing we have discovered is that Obama's performance among uneducated whites in Democratic primaries anywhere is approximately 25%-30%.  That's going to take a lot of work to overcome, and I really hope that we can rely on surrogates like Edwards or the Clintons to really campaign hard in Appalachia and the white parts of the Deep South to convince them that the message is the right one for them (economics and the war will be our strongest weapons here).

by Dracomicron 2008-05-16 12:04PM | 0 recs
Re: Great diary

Appalachia is a dead end for us now...  The coastal South is not, though...

by LordMike 2008-05-16 12:42PM | 0 recs
Re: Great diary

Larry Sabato has an article talking about the uneducated blue collar vote.  Basically, he wonders what the hubaloo is all about?  The numbers are smaller than the media says, and they almost always go republican anyways...  He feels that this is a silly demographic to pander to.  Much more can be gained elsewhere.

Sabato knows what he's talking about!  He's been spot on in all the elections this century.

by LordMike 2008-05-16 12:45PM | 0 recs
I wouldn't pander to them, but...

If we set Bill Clinton on that demographic to talk up Obama's agenda, then it would sure make McCain sweat.

by Dracomicron 2008-05-16 12:50PM | 0 recs
Re: I wouldn't pander to them, but...

Doubtful... he did very little to nothing for Hillary in NC.

by LordMike 2008-05-16 12:53PM | 0 recs
NC was a lost cause

Seriously, there was nothing they could've done there.

On the other hand, a ton of West Virginians said that Bill touring there was important to them.

by Dracomicron 2008-05-16 01:01PM | 0 recs


Advertise Blogads