Articles and Analysis


Undecided Voters and Racial Attitudes


How will undecided voters break, and will racial attitudes color their votes?

We've seen an enormous amount of speculation but little evidence based on data, so let's try to tip the balance back to empirical evidence.

Thanks to the Diageo/Hotline tracking poll data, we can model individual vote choice and see what we would expect of undecided voters.

During October 3-11 our colleagues at the Diageo/Hotline poll included a racial attitude question we had previously used in the Big Ten Battleground survey in September and which NBC/Wall Street Journal used in January. That question was shown in both the earlier polls to have a statistically significant effect on vote choice, even after controlling for other political attitudes and demographics.

The question text is: "I'd like you to tell me whether you strongly agree, somewhat agree, somewhat disagree or strongly disagree with the following statement. ... African Americans often use race as an excuse to justify wrongdoing"

58% of the total sample, and 61% of whites agree either strongly or somewhat. (For comparison, 56% of hispanics and 40% of African Americans agree strongly or somewhat.)

I model the vote choice for those who expressed a preference with a model using a variety of attitudes and demographics, including favorability to Obama and McCain, party id, marital status, kids at home, education, race, age, sex, church attendance, region and urban, suburban or rural residence. Then I added the racial attitude responses from the "black excuse" question. To check against people hiding their feelings by refusing to answer the "black excuse" question I also included a variable to capture the effect of refusal to answer.

This model produces a predicted probability of voting for McCain or Obama, including predicted probabilities for those who had said they were undecided or who refused to respond to the vote question. From this we can estimate the likely vote of undecided, and compare the estimates to the responses of those who gave a vote preference in the survey.

Bottom line: Undecided and refuse to say voters are estimated to break 50% for McCain and 50% for Obama. As even as it gets. There is no evidence here of a large bias towards McCain that is hidden within the undecided respondents.

Nor is there evidence of a pronounced racial bias among these undecided voters as compared to the public at large. Among the undecided 27% strongly agree and 32% somewhat agree on the "black excuse" item. For the public as a whole 26% and 32% give the corresponding responses.

The model does a good job predicting survey response as well. 97% of both Democratic and Republican voters are predicted by the model to vote that way. For those who say they only "lean" towards one party or the other, 77% of Democratic leaners and 80% of Republican leaners are predicted to vote as they lean. The symmetry of results here suggests that there is not a visible bias in the model estimates for either party or for intensity of preference.

Finally, what happens if we ignore racial attitude and predict vote among the undecided without it? The split is 52% Obama to 48% McCain. So at most the impact of incorporating racial attitude in the model is a rise of 2% for McCain among undecided. Given the sample sizes involved, that is well within the margin of error. And if we take out candidate favorability from the model we get estimates of 52-48 without racial attitude and 53-47 with racial attitude.

So what can we conclude? There is no evidence of a hidden support for McCain among undecided voters. They split more evenly than does the "decided" pool of respondents, who split 54-46 in this sample (Oct 3-11) but that's well within normal expectations and is a modest difference in any case.

Second, the role of racial attitude is important at the individual level, but the aggregate consequence is extremely modest. Some are moved away from Obama yet others are moved towards him. And among the undecided, the distribution of opinion on this measure of racial attitude is virtually identical to that in the population.

In a year of endless discussion about racial effects there has been far more speculation and far less data analysis than is good for us. Let's put our data on the table before continuing to opine about this subject.



Charles, this is an excellent little study. The results are useful to those who might expect the uncommitted to break toward Obama or McCain based on racial attitudes.

The results are also suggestive of the idea that there isn't a hidden "Bradley Effect" in the undeclared/uncommitted respondents.

However, there still could be "hidden" racial bias among nonrespondents -- those who did not participate in the survey. And this wouldn't be captured in the survey at all.

Thanks very much for your work on this topic.



The results are also suggestive of the idea that there isn't a hidden "Bradley Effect" in the undeclared/uncommitted respondents.

Not to get too awful nit-picky, but the Bradley effect, if there is or ever was one, is by definition hidden. Bradley's and Wilder's support were not overestimated by the polling. Their opponents' support was underestimated, and hidden in large numbers of undecideds. At least that's the way the story goes. Or went.



The McCain campaign released a memo last night indicating they expected the undecideds to break almost lock-step their way. In an interview this morning, they indicated that their data showed that the undecideds were elderly, rural, and white, which they exptected to win handily. Also, during the primaries, the "late deciders" broke against Obama by about a 70-30 margin. I'd be curious as to what you thought about it.



Prof Franklin,

Thanks so much. An excellent analysis and very good empirical refutation based on a question with excellent distributional characteristics.



I can't thank you enough for this--it really irks me that some people opine with great confidence and yet so little empirical basis on an issue as important as the effects of race on politics.



Thank you for the analysis... I was getting sick of the "undecideds will definitively break for McCain" argument bouncing around the Internet which had little basis in actual analysis but a lot of "out of the ass" assumptions...




Re the McCain campaign spinning: what else do you expect them to say? No public polls back up their argument.

Re the Dem primary: it was Dem vs. Dem then and has little relation to a general election. Applying primary election trends to a general election is putting a square peg in a round hole. But in an effort to compare something that inherently incomparable: Hillary was a known factor and people with little information will naturally vote for a name they know; yes, Obama had been campaigning for a year but he had spent most of that time camped in Iowa. The primaries ended nearly 6 months ago and, during that time, Obama has made himself known to the entire American populace.


I agree with the previous comment suggesting that this problem is, by definition, unmeasurable. Or at least very difficult to measure with questions such as this one.

An excellent effort on Charles' part, but I have to questions how many hardcore racists would answer this racial attitude question honestly.



This does show a small shift, even if it is within the MOE. But Bradley-Effect or not, people are not going to vote for Obama because he is black. And they'll say it up front. I think an interesting study would be to see if black candidates nationwide fair better after two years of an Obama presidency. Possibly a comfort factor that would win over those working-class whites. All those illusions that McCain is trying to make stick to him would prove false and that alone will be a big help to race relations in the US.



Just an observation that the hard-core racists appear to have no qualms about speaking their views - I've been pretty shocked how willing people have been to speak to television crews very openly about how they would never vote for a black candidate. It's also fair to assume that the hard-core racists were never going to vote for Obama anyhow and are therefore a small part of the undecided vote.



Sometime in September I was polled with a question similar to the one described above. I immediately hung up on the pollster finding the question offensive (I am white). This was asked before the demographic questions were asked. I think that instead of a "Bradley effect" there may be an undercounting of people who did not complete the survey due to the question itself.



Prof. Franklin - This may show that I am obtuse, but I want to be sure I'm clear on what you did. I think what you are saying you did (and maybe it's totally clear to everyone else) is construct a model that would predict voter choice and then tested it among the voters who expressed a definite preference. In that test, the model correctly predicted the choice of 97 % of those who expressed a choice. You then used that same model among the leaners and undecided and got the results described above. Is that correct? If not, can you clarify?


Professor T:

Many of the undecided respondents won't break at all. They won't vote.

Many of the polls are using expanded LV models, and for good reason. But this will result in overinclusive results. A lot of disengaged and apathetic respondents will not admit as much, and will instead claim that they intend to vote. It seems to follow, in my mind at least, that these "voters" are more likely to be undecided. It also follows that these "voters" are much more likely to not actually vote.

I know that comparisons of the polls done with the various LV models don't seem to show big differences in their results, but I still think that the collective impact of the more inclusive universes being used this year will mean that more undecided respondents will not vote.

If my theory is even partically correct (e.g., if a higher percentage of undecideds don't vote, as compared to the decided respondents), then the undecided pool will have reduced impact on the final results. Since McCain seems to need all the undecided and then some, this could be yet one more reason his chances are extremely poor.


Professor Franklin:

The one quibble I have is that you did not split up the voters who had expressed a preference into a training set and a validation set to test your fit. At this point I cannot be sure you fit is not artefactual, though I admit that is unlikely. I just think it is much cleaner to check your fit on a completely different set of data.

Otherwise, I think it's quite interesting and provocative.


Vicente Duque:

Racism will be (temporarily) defeated on November 4 :

There are many clues, indications and intimations that show that Obama already won this election before November 4.

But the most powerful of all evidencies is Mathematics.

Since Obama is so many points ahead in the most prestigious polls. And the Undecideds are a small proportion, then even adding all the undecideds McCain can not win.

There are past elections that indicate that Undecideds have a tendency to split evenly at the last moment, and not by big margins.

INTRADE has predicted correctly several elections. Bettors have "insider information" and "intimate information", they know more than the regular person, or common person in the street.

Bettors are inside campaigns, or have political and campaign experience, they know the state where they live.

The can't be wrong with their money.

And the fact that just a few states decide all conspires against McCain. It is easier for Obama to control a few states like Ohio or Florida. The Electoral College system does not help McCain.

And the big amounts of Money and Advertising also signal doom for McCain.

There are no more Television Debates where McCain can take advantage of Obama. And Americans are not so stupid as to vote because of Racial Hatred. If your house is burning, then the racial color of the Firefighters does not matter.

Last but not least, Minorities in Florida and the Southwest ( Colorado, Nevada, New Mexico ) have been eager to vote for Obama in larger turnout and long lines. That has been reported by journalists and TV. They can be decisive, specially in the mentioned States.



Vicente Duque


Post a comment

Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.