Racial attitudes have not improved in the four years since the US elected its first black president, an Associated Press poll has found.
Released on Saturday, the AP poll used a combination of explicit and implicit questions about race and found that prejudice has increased slightly since 2008.
The study, conducted online by GfK Custom Research under the supervision of AP’s polling unit, included interviews with 1,071 adults between August and September 2012.
"As much as we'd hope the impact of race would decline over time ... it appears the impact of anti-black sentiment on voting is about the same as it was four years ago"
- Jon Krosnick, Stanford University
In all, 51 per cent of those polled expressed explicit anti-black attitudes, compared with 48 percent in a similar 2008 survey.
When measured by an implicit racial attitudes test, the number of Americans with anti-black sentiments jumped to 56 per cent, up from 49 per cent during the last presidential election.
In both tests, the share of Americans expressing pro-black attitudes fell.
"As much as we'd hope the impact of race would decline over time ... it appears the impact of anti-black sentiment on voting is about the same as it was four years ago,'' said Jon Krosnick, a Stanford University professor who worked with AP to develop the survey.
Most Americans expressed anti-Hispanic sentiments, too. In an AP survey done in 2011, 52 per cent of non-Hispanic whites expressed anti-Hispanic attitudes. That figure rose to 57 per cent in the implicit test.
Experts on race said they were not surprised by the findings.
"We have this false idea that there is uniformity in progress and that things change in one big step. That is not the way history has worked,'' said Jelani Cobb, director of the Institute for African-American Studies at the University of Connecticut.
"When we've seen progress, we've also seen backlash.''
Obama himself has tread cautiously on the subject of race, but many African-Americans have talked openly about perceived antagonism toward them since Obama took office.
As evidence, they point to events involving police brutality or cite bumper stickers, cartoons and protest posters that mock the president as a lion or a monkey, or lynch him in effigy.
"Part of it is growing polarisation within American society,'' Fredrick Harris, director of the Institute for Research in African-American Studies at Columbia University, said.
The poll finds that racial prejudice is not limited to one political group. Although Republicans were more likely than Democrats to express racial prejudice in the questions measuring explicit racism (79 per cent among Republicans compared with 32 per cent among Democrats), the implicit test found little difference between the two parties.
That test showed a majority of both Democrats and Republicans held anti-black feelings (55 per cent of Democrats and 64 per cent of Republicans), as did about half of political independents (49 per cent).
|In-depth coverage of the US presidential election
Obama faced a similar situation in 2008, the survey then found. The Associated Press developed the surveys to measure sensitive racial views in several ways and repeated those studies several times between 2008 and 2012.
The explicit racism measures asked respondents whether they agreed or disagreed with a series of statements about black and Hispanic people.
In addition, the surveys asked how well respondents thought certain words, such as "friendly'' "hardworking", "violent" and "lazy", described blacks, whites and Hispanics.
The same respondents were also administered a survey designed to measure implicit racism, in which a photo of a black, Hispanic or white male flashed on the screen before a neutral image of a Chinese character.
The respondents were then asked to rate their feelings toward the Chinese character.
Previous research has shown that people transfer their feelings about the photo onto the character, allowing researchers to measure racist feelings even if a respondent does not acknowledge them.
Results from those questions were analysed with poll takers' ages, partisan beliefs, views on Obama and Romney and other factors, which allowed researchers to predict the likelihood that people would vote for either candidate.
Those models were then used to estimate the net impact of each factor on the candidates' support.
Previous studies have shown that poll takers are more likely to share unpopular attitudes when they are filling out a survey using a computer, rather than speaking with an interviewer.