Survey Bias on the Front Porch

Abstract

We examine whether survey interviewers are biased in their views of certain classes of respondents, thereby introducing unobserved bias into survey results. There has been a great deal of previous research on how racial and gender dynamics affect the responses given by respondents during face-to-face surveys. In this article, we turn this issue around and ask whether human interaction affects how the interviewer views the respondents, and if so, how this may systematically bias surveys. If interviewers are biased, this may impede their ability to conduct interviews in a consistent, nonjudgmental, and unbiased manner. Using three surveys that required the interviewer to evaluate how informed and intelligent the respondents appeared, we found that interviewers were more likely to evaluate respondents of lower socioeconomic status as less informed and less intelligent, even after controlling for objective levels of political information. There is also evidence that Blacks may be negatively evaluated.