View article

In this paper, we expand the existing research on interviewer effects in telephone surveys by investigating how interviewees respond differently to interviewers from differing racial backgrounds, leveraging an experimental design based on an original telephone survey of African American southerners conducted in February 2009, shortly after the election of Barack Obama as president in which respondents were randomly assigned to white or black interviewers. We find, using logit models, that African American southerners offer different responses to white interviewers than black interviewers when prompted with racially-valenced questions, despite substantial political and social change over recent decades, although these differences do not emerge in questions that do not implicate race directly.

Previous research has demonstrated that survey participants may alter responses given to an interviewer of a different race (see e. g. Hyman et al. 1954; Schuman and Converse 1971; Davis 1997a, b; Anderson et al. 1988b). This phenomenon has often been most significant when a survey of an African American respondent is being administered by a white interviewer (Davis and Silver 2003). However, just as well established phenomena—such as racism among southern white males—continue, they also evolve in subtly important and measurable ways (Kuklinski et al. 1997). Does the uniformity in the difference in answers given by African American respondents to white interviewers still exist, or has a new pattern emerged? While pundits who were quick to herald a “post-racial America” in the wake of Barack Obama’s election have been brought back …