Doctoral Dissertation Research: Gaze Patterns During Video-mediated Interviews

The New School, New York NY

Investigators

Abstract

This research project will explore how and why the small self-view window (a window with a live video feed of oneself in the corner of the screen) in video-mediated communication (such as Skype) can affect people's disclosure of sensitive information during survey interviews. In an earlier study, respondents in the self-view condition disclosed more socially undesirable (that is, more embarrassing) behaviors and perceived the interview to be less sensitive. This project will expand on earlier research by tracking where survey respondents look on the screen while answering sensitive questions via video. Examining gaze patterns and how they might differ for different respondents or questions will provide a basis for understanding why people disclose more and find the interview less sensitive with a self-view window. The findings will increase understanding about how nonverbal behaviors correlate with disclosure. Video-mediated interviewing is a potentially cost-effective mode of data collection that could facilitate the collection of policy-relevant data on some hard-to-reach populations. The data collected may be useful for informing the design of new video-based data-collection interfaces. Project results will be presented at conferences and publication venues that maximize impact across several disciplines. As a Doctoral Dissertation Research Improvement award, support is provided to enable a promising student to establish a strong, independent research career. Video-mediated interviewing offers a unique combination of features that differ from those in other face-to-face or audio interactions, including the ability to see oneself in a self-view window. This research project will compare the responses and gaze patterns of adult participants interviewed in a controlled laboratory setting over Skype. The experiment will use a set of sensitive and non-sensitive questions from U.S. government and social scientific surveys. Respondents will be randomly assigned either to a self-view (respondents can see themselves and the interviewer) or no self-view (respondents can see only the interviewer) condition. Measures of gaze will be recorded using an unobtrusive eye-tracking system. The data will be analyzed to determine when and how often self-view respondents look at the images of themselves, whether respondents who disclose more look at the self-view more, and how gaze location, duration, and frequency of glances differ for sensitive vs. neutral questions and for less socially desirable answers.