a PSC Research Project [ARCHIVE DISPLAY]
Investigator: Frederick G. Conrad
The advent of new technologies such as animated or conversational agents has the potential to merge important benefits of interviewing (e.g. higher response rates, improved respondent comprehension) with those of self-administration of survey questionnaires (lower cost, increased privacy, convenience and control for respondents). Animated agents are software objects in the user interface that can range from simple line figures (like the widely detested Clippy) to much more successful embodied avatars that gesture, display facial expressions and move their eyes in coordination with their speech. It seems almost certain that developers of web surveys will explore animated agents as a way to combine the benefits of interviewing and self-administration. We propose a series of laboratory experiments to examine the impact of different features of agency on responses to sensitive and non-sensitive survey questions. Rather than developing agent software we will simulate agents using a “Wizard-Oz-Technique” in which respondents believe they are interacting with a computer generated agent when in fact they are interacting with an actual interviewer whose image has been digitized and rendered graphically in the user interface as if it were computer generated. The overall goal of the proposed work is to determine when animated agents might help and when they might hurt quality of responses.
|Funding:||National Science Foundation (SES 0551300)|
Funding Period: 09/01/2005 to 09/30/2010
This PSC Archive record is displayed for historical reference.