Given a choice of several positive and negative adjectives that might describe blacks, 20 percent of all whites said the word "violent" strongly applied. Among other words, 22 percent agreed with "boastful," 29 percent "complaining," 13 percent "lazy" and 11 percent "irresponsible." When asked about positive adjectives, whites were more likely to stay on the fence than give a strongly positive assessment.

Every month or so, polls come out giving approval ratings of "politicians", "journalists", etc.  These are always really, really, low.  I have no doubt that everyone has a few varying degrees of bias (racial, cultural, gender, xenohobic, anti-etc.), but I never quite understand the media reporting on those results.  I have similar reservations about this poll, but maybe that's just me.

The AP's analysis seems to be pretty good here, but it jumps around a lot between taking a balanced view of the big picture (like this)...

Race is not the biggest factor driving Democrats and independents away from Obama. Doubts about his competency loom even larger, the poll indicates. More than a quarter of all Democrats expressed doubt that Obama can bring about the change they want, and they are likely to vote against him because of that.

...and dropping media buzzbits like this...

"We still don't like black people," said John Clouse, 57, reflecting the sentiments of his pals gathered at a coffee shop in Somerset, Ohio.

Among white Democrats, Clinton supporters were nearly twice as likely as Obama backers to say at least one negative adjective described blacks well, a finding that suggests many of her supporters in the primaries -- particularly whites with high school education or less -- were motivated in part by racial attitudes.

An election, and especially a primary, is not an either/or selection, or even (though the conflation of market research and opinion research  would prefer otherwise) that choosing a brand reflects pyschographic characteristics irrespective of the material product itself.

A voter can be highly motivated for positive reasons to select a candidate, and also not be averse to another candidate due to underlying prejudices. As one of the above links note, something like 58% of those with negative responses (in this survey) about "blacks" indicate an intention to vote for Obama (and the premise that they will not act on that stated intention, the Bradley effect, is refuted by Giordano, Silver, et al.).

Lastly, which response(s) is (are) symptomatic of racism? If you look at the results, question RAC14, on page 23, 84% of all respondents and 90% of all whites think that "blacks" are responsible for at least "some" of the racial tension in the United States. That can be read a couple of different ways, but after 300 years of slavery and Jim Crow, African-Americans are responsible for "some" of the racial tension???  If you consider that question a measure of racism, then we are all racists now... and we are voting for Obama by a wide margin.  

But I don't understand this post, J.  Is it "Racism!" mongering, or is it just "race factors may or may not hurt Obama.  Hmmm.  Intersesting."

I can comment on the former; the latter just leads to "what is it I am allowed to say?

You can poll-test for a host of social identity metric.  Any.  Race, Gender, Age.

I don't see why "Race" is now the qualifier.  It sounds like you're buttressing the cushion that "If Obama loses, it was RACISM!"  

Something we decried in the Primaries when we were on Hillary's side.

1.  More likely/less likely means nothing unless it CHANGES your vote.

2.  Even if 20% of Hillary supporters really vote for McCain, that would even be less than 10% of all Democrats (20% times her 50% of primary vote, and remember some of those were I's and R's).  Even Bush got 13% of Democrats vs. Kerry.

3.  Oh yeah, none fo the polls are showing that kind of effect anyway.

Real Clear Politics Poll of electoral votes

The AP did all its critical musical chairs shifts in this past year - in preparation for this election coverage...