Racism and Sexism on Super Tuesday

Barack Obama squeaked out a narrow win on Super Tuesday to expand his small lead in the pledged delegates. While many of the exit polls were off as predictions, these polls do provide us with one of the best national tests of to what extent voters will admit to racist or sexist motivations. And they also tell us an important fact: Obama was hurt by race-based voting, while Clinton was greatly helped by gender-based voting.

More Democratic voters admitted to racist motivations in opposing Obama (an average of 2.88% of the voters) than admitted to sexist motivations in opposing Clinton (an average of 1.83%). Overall, the racist vote outweighed the pro-black vote for Obama by an average of 0.5%. By contrast, the pro-woman vote for Clinton outweighed the sexist vote by an average of 5.0%. That’s a 5.5% swing in each state. The continuing significance of racism (beyond the expressed levels of sexism) was a key factor in why Obama didn’t win a decisive victory over Clinton on Tuesday.

The Democratic exit polls asked voters, “In deciding your vote for president today, was the gender[race] of the candidate a) The single most important factor; b) One of several important factor; c) Not an important factor.”

The race-based voting data is summarized online in a spreadsheet I created. The gender-based voting data is in another spreadsheet of mine.

Here is the average of 15 states and the views on race of those opposing Obama (note, this is an average of the states, not an average nationwide):
2.87% Super racist (race is “the single most important factor,” voted against Obama)
6.61% Partly racist (race is “one of several important factors,” voted against Obama)
9.49% Total of racist vote (super racist plus partly racist)

Here’s is the average of the voting based on race by Obama supporters:
2.0% Super pro-black (race is “the single most important factor,” voted for Obama)
6.99% Partly pro-black (race is “one of several important factors,” voted for Obama)
8.98% Total of pro-black vote (super pro-black plus partly pro-black)

Here is the average of the 15 states with exit polls on their response to the question of voting based, on gender, against the female candidate:
1.83% Super sexist (gender is “the single most important factor,” voted against Clinton)
6.08% Partly sexist (gender is “one of several important factors,” voted against Clinton)
7.9% Total of sexist vote (super sexist plus partly sexist)

Here is the average of the 15 states with the pro-female voting:
2.66% Super pro-women (gender is “the single most important factor,” voted for Clinton)
10.27% Partly pro-women (gender is “one of several important factors,” voted for Clinton)
12.93% Total of pro-women vote (super pro-women plus partly pro-women)

I checked to see if the average admitted racism might have been a factor in the states with big discrepancies between exit polls and final results, but I don’t see anything. This doesn’t mean that a Bradley effect is impossible, but we’ve seen polling misses in both directions in 2008, and much of Clinton’s disproportionate advantage over the exit polls may have come from early voting.

Some might say that I’m applying a racial double standard. That’s because I am. I’m saying that it’s perfectly fine for black voters to say that they support the black candidate. Partly that’s a rational belief that a black candidate (with strong ties to the black community) will represent the interests of black people. But it’s also a double standard based on the history of racism in America, because the NAACP is not the equivalent of the KKK, even if both promote a certain kind of racial uplift. The historic oppression of black people makes it quite rational for everyone (of any race) to care about electing a black president). By contrast, in a country of historic white racism, it is racist to oppose a black candidate based on race. I apply the same reasoning to gender: it’s reasonable for women (and men) to support electing the first female president, and it’s vile sexism for men (or women) to oppose a woman president based on her gender.

But in general, gender-based voting helped Hillary Clinton. By contrast, race-based voting usually hurt Barack Obama. That suggests racism is a more powerful negative force than sexism in this election. I would assume that voters in exit polls are more likely to admit sexism rather than racism, because there is a greater social stigma toward admitting racism, and therefore more racism against Obama may have been concealed.

Although we would expect that white voters reflect the vast majority of this racist voting, we should not ignore other racial groups. Are Latinos racist against black politicians? Certainly, the massive Latino support for Hillary Clinton is troubling, most especially because Obama has a more progressive position on illegal immigrants, such as his support for driver’s licenses. Some of the support for Clinton may reflect the institutional bias toward her, and the difficulty in reaching Latinos with alternative news sources. In California, Asians were the racial group most supportive of Hillary Clinton (71%), even more so than Latinos (67%). Why is that so shocking? Perhaps it’s because Barack Obama is the first presidential candidate who grew up for part of his youth in Asia, with an Asian sister and an Asian stepfather. Could it be the case that a small (but significant) percentage of Asian voters are so deeply prejudiced against black men that they would oppose the candidate with the deepest roots in the Asian community of any presidential candidate in history?

It is even possible that African Americans are voting against Obama based on race, although given their strong support for him (around 80%) this seems a minor factor at most. But Democrats of all races may be voting against Obama based on race even though they are not racists. Instead, they imagine that Americans are racist and won’t elect a black president, they fear losing in 2008, and therefore they think that a black candidate must be defeated. This is a very odd sort of reasoning (since all polls and reasonable analysis indicate that Obama is more electable than Clinton), but it could be a factor. Is there some non-racist reason why someone would vote for the white candidate because of race? None I can think of. Even the “rational racism” I describe above is just another form of racism.

What should be the reaction to this information? First, we must acknowledge the persistence of racism (and sexism), even among the Democratic voters. I don’t necessarily think, however, that this small group of Democratic voters will turn to the Republicans due to bigotry (a few bigoted voters, especially against Hillary Clinton, may be Republicans voting in the Democratic primary and willing to express their sexism). What’s important is to speak out against this bigotry, and to encourage voters of all races to reconsider their opposition to a black or female candidate. Most of all, these figures should be used to mobilize true progressives to get out the vote and overcome the remaining vestiges of bigotry in America.

(A caveat on the data. This is complicated, so corrections are welcomed. Because I do not have raw numbers, I used the MSNBC reported exit poll figures, which are rounded to the nearest number. As a result, a common “3%” figure could be 2.5% to 3.49%, which allows for a lot of variability. In some cases, the strong declarers are such a small number that the exit polls did not report who these people voted for. In those cases, I calculated the figure by taking the overall “important” number reported and determined how the “single most important factor” votes must have been by comparing them with the “one of several important factors” figure. Once again, this can create a fair amount of variability due to rounding and small sample sizes. I didn’t weight the states by size in my average, but somebody could; I suspect it wouldn’t have a dramatic effect. So the overall averages should be pretty accurate, but individual states could be off target. I believe that Edison Media Research and Mitofsky International conduct the exit polls, but I don’t know the sample sizes. If anybody has any raw data about this, I’d love to see it.)

Crossposted at DailyKos.