Virtual Interviewers, Social Identities, and Survey Measurement Error
Conrad, Frederick G., Michael F. Schober, Daniel Nielsen, Heidi Ann Reichert, Michael F. Schober, Daniel Nielsen, and Heidi Ann Reichert. 2020. "Virtual Interviewers, Social Identities, and Survey Measurement Error." In Interviewer Effects from a Total Survey Error Perspective. Chapman and Hall/CRC.
This chapter explores an approach that may incorporate attractive features of both interviewer- and self-administration while reducing monetary costs and measurement error inherent in either approach. The proposal is that by implementing embodied, animated agents or "virtual interviewers" (VIs), which are considerably cheaper than live human interviewers, it may be possible to engage respondents more than in traditional text-based web surveys, while deliberately using VIs' social identities to improve the quality of answers. In general, the attributes of VIs are largely under the control of researchers and designers. While interviewer effects can undoubtedly harm data quality, interviewers' social identities may also be exploited to improve data quality. When offered a choice of VI for a hypothetical future interview, Black respondents primarily chose Black VIs, perhaps increasing their comfort. Voice was frequently reported as a criterion for choosing a VI even though voice characteristics were not systemically varied.