Sanders: Perceptions of race affect survey responses

In reviewing the National Black Politics Study -- a phone survey of black Americans conducted entirely by black interviewers -- Lynn Sanders, Assistant Professor in Political Science, found that not only did respondents attempt to determine the race of the interviewer, but they clearly adjusted their responses based upon what they perceived the caller's race to be. Respondents who believed the interviewer was white provided more moderate opinions than those who thought they were sharing their opinions with a fellow African American, she found. Other research has shown that whites respond the same way -- moderating their views -- when talking with blacks.

"Racial attributions are such pervasive and entrenched elements of political life in the United States that they affect public opinion even when people of the same race are talking to each other on the telephone," said Sanders, who discusses her work in the paper "What Is Whiteness?"

The survey was made of 1,200 African Americans who were asked about political and cultural issues. Overall, the survey found a shift in attitudes toward more support of black nationalism.

The opinion study, which was conducted from November 1993 to February 1994, was funded by the Russell Sage Foundation and conducted by Market Strategies Inc., a polling firm in Southfield, Mich.

Michael Dawson, Associate Professor in Political Science, and Ron Brown of Wayne State University are principal investigators for the survey, which has become an important source of data for researchers studying trends in political opinions among blacks.

Perceptions of race reveal view of power

The research firm selected black interviewers because it felt white interviewers would have a more difficult time gathering information about politics and other issues from black respondents. At the end of the interview, the people being questioned were asked if they thought the interviewer was black or white. In 75 percent of the cases, the respondent was certain that the interviewer was black, while in 14 percent of the cases, the interviewer was identified as white.

Sanders suggests that in assessing an interviewer's race, the respondents may have relied on such factors as the interviewer's speech or grammatical patterns, the tone or pitch of the interviewer's voice, the humor he or she may have used and the name given by the interviewer at the beginning of the call.

When Sanders looked at the responses people gave, she found a marked difference between the way blacks talked to other blacks and the way they talked to people they assumed to be white. They tended to present more centrist views when they believed the person they were talking to was white.

When asked if they thought blacks could be best represented politically by blacks, for instance, 64 percent of the respondents who thought they were talking with other blacks said they agreed; among the respondents who thought their interviewer was white, 49 percent agreed.

When they were asked their views on the statement "American society is unfair to blacks," 83 percent said they agreed with that view when they thought they were being interviewed by another African American, while only 67 percent of those who perceived their interviewers to be white said they held that view.

In speaking with interviewers they believed to be white, respondents were more critical of Louis Farrakhan and rap music than they were when they were talking with people they believed to be fellow African Americans. When asked if they felt that rap music was a destructive force in the black community, 50 percent of those who felt they were being interviewed by blacks agreed, while 65 percent agreed of those who thought their questioners were white. When asked if they thought Louis Farrakhan was a dangerous extremist, 28 percent of the respondents agreed with the view when they thought they were talking with fellow blacks, while 40 percent agreed when they thought they were talking with whites.

The responses reveal the way in which blacks think power in the United States is associated with race, Sanders said.

"Whiteness seems unequivocally associated with something that might roughly be summarized as state power," she writes. "Respondents generating opinions on the legitimacy and fairness of state-sanctioned political institutions and practices appeared to make an assessment of the race of the person to whom they addressed these opinions. These results suggest that the institutions of the American state are far from race-neutral in the eyes of black American citizens."

Consequences of the findings

Researchers have long examined the impact of interviewers on the outcome of surveys. That research has led some scholars to feel that blacks are more willing to reveal their true attitudes when they are interviewed by African Americans.

The impact of this accommodation to a perception of race is important to people who follow elections, as polls are frequently used to assess a candidate's chances of being elected. If people adjust their responses based on their determination of an interviewer's race, then polls projecting the potential appeal of a black candidate may be flawed if blacks are being interviewed by people they may think to be white. The black respondents may be reluctant to share their support for other blacks with a white interviewer and may instead offer what they consider a conciliatory viewpoint.

The study further shows how deeply divided by race America is and points to contradictions apparent in political discussions.

"As much as Americans of all races are highly conscious of race, they are also deeply devoted to a wish not to notice race," Sanders said. "A very long strain of critical thinking about U.S. politics insists that race should and might be irrelevant." People who suggest that reverse discrimination exists in America and who oppose affirmative-action programs often do so with a claim that race should not matter, she added.

"My analyses showing the power of mere perceptions of race do not speak directly to the question of whether race should be irrelevant in American politics. They say a great deal, however, about whether it might be," she said.

"The results from the National Black Politics Study suggest that it is highly unlikely that ordinary citizens might conveniently forget to think in racial categories. These categories are entrenched and are utilized by Americans in systematic ways as they state their opinions on a variety of political issues."

-- William Harms