Association for Public Policy Analysis & Management

Seminal texts in public opinion research have addressed such issues as “the American voter,” “the changing American voter,” the “unchanging American voter,” and “the vanishing voter.” This most recent election cycle brought to light another kind of voter – one whose self-reported electoral choices may have been influenced more by a heightened climate of race and gender issues instead of a true depiction of how they actually felt. This type of voter may have been reluctant to express certain attitudes or may have chosen to answer questions in ways deemed socially acceptable … all because of the voice at the other end of the phone. This effect is hypothesized to be correlated with respondents’ choices for both candidates and policy preferences.

Past interviewer effects research has addressed respondent interaction with interviewers of certain races and of a certain gender, and the interviewer effects research specifically on candidate choice has focused on the social and racial identity of the candidates in question themselves – such as the Wilder Effect, the Bradley Effect, and the Whitman Effect. But the unprecedented nature of the 2016 election cycle had unprecedented consequences for interviewer-respondent interaction, especially regarding President-elect Donald Trump.

Throughout the course of the general election campaign and now post-election period, we explored gender and race of interviewer effects using a diverse pool of telephone survey interviewers, mainly drawn from the majority-minority Rutgers University student population. We assessed reported 2016 attitudes by gender, as well as by both the interviewer’s actual race and the interviewer’s race as perceived by the respondent. We then compared reported attitudes in live telephone interviews to those done through text and email at a similar time.

This research has important implications for interviewer effects, interviewer populations, as well as social desirability bias that extends to both expressed candidate preferences and policy positions. In addition to examining interviewer effects on candidate choice, we also measured whether interviewer effects are correlated to expressed policy preferences expressed by survey respondents in four policy areas: immigration policy, civil rights, gender equality and Affordable Care Act (Obamacare).