There is pressure for large surveys to move to mixed mode data collection. Switching modes across waves of longitudinal study or using different modes across subgroups may result in non-comparable results, and therefore a better understanding of mode effect sources is required. We present the first study of the feasibility of tracing respondents’ eye-movements in order to compare them across modes, including web, self-completion paper and pencil (SAQ) and face-to-face with show cards. Observing the eye-movements of 16 respondents interviewed in a lab, we find that respondents often don’t read questions and response options in the sequential order, and this differs substantially between face-to-face and self-completion modes. By studying the influences of different reading styles on responses, we find a suggestion that satisficing (superficial answering) may lead to socially desirable responses in self-completion, but not in face-to-face mode. We test this hypothesis in a separate lab study with a 2×2 design where motivation and response order are manipulated. We find that motivating respondents to answer more thoughtfully leads to more revealing, socially undesirable answers. This suggests new ways for survey methodologists to improve responses to sensitive questions. We conclude with discussion of the potential of eye-tracking in understanding and improving response quality across modes.
Presented by:
Olena Kaminska (ISER)
Date & time:
May 29, 2013 12:00 pm
Internal seminars home