Gender-of-Voice Effects in an ACASI Study of Same-Sex Behavior
Fahrney, K. M., Uhrig, J. D., & Wolitski, R. (2005, May). Gender-of-Voice Effects in an ACASI Study of Same-Sex Behavior. Presented at American Association for Public Opinion Research 60th Annual Conference, Miami Beach, FL.
Audio computer assisted self interviewing (audio-CASI or ACASI)has become an increasingly popular method for administering surveys ofsex and other sensitive behaviors. One of the major advantages of ACASIis that it removes the requirement that respondents divulge sensitivebehaviors directly to another person. As a result, ACASI may reduce theextent to which response accuracy is compromised by interviewereffects. However, the literature on Computers As Social Actors (CASA)suggests that computer users ascribe human traits to computers; evensubtle cues such as the gender of the computer-generated voice maycause users to react to the computer as they would to another person.Based on these findings, we could infer that there may be the samepattern of gender-of-interviewer effects when using ACASI techniques asthere are in traditional face-to-face interviewing. If this were true,the choice of an ACASI voice may be just as important as the choice ofinterviewing staff in traditional face-to-face interviewing. Althoughthere have been a few studies of the effects of gender-of-voice intelephone-ACASI and IVR, there is no published research ongender-of-voice effects in ACASI studies. The present study examineswhether the gender of the ACASI voice impacts the reporting of sociallyundesirable HIV-risk behaviors in a sample of men who have sex withmen. A sample of 405 respondents were randomly assigned to the male orfemale voice condition. Our preliminary results suggest that a malevoice yielded more socially desirable responses than a female voice toquestions about the number of partners and unprotected sex withpartners with unknown HIV status. We discuss our results in terms ofsocial attribution and social distance theories.