Race-of-interviewer effects and survey questions about police violence

ABSTRACT

This research uses binary logistic regression to test for a connection between the race of interviewer and race of respondent on five questions in the General Social Survey about the use of physical force by the police. Results indicate two instances of race-of-interviewer effect: (1) black respondents were more likely to voice disapproval about whether the police can strike a citizen trying to escape when speaking to a black interviewer, and (2) white respondents were less likely to voice approval of police striking an adult male citizen in the presence of a black interviewer. Secondary findings indicate that education is consistently significant regardless of race of respondent and the survey question, while social class, sex, age, and region are significant in only limited scenarios.