![IconGems-Print](images/IconGems-Print.png)
Implications of Response Device Type for Sensitive Web Surveys: Examining Data Quality and Respondent Characteristics in a Survey of College Students
Amanda Smith
RTI International
Ashley Richards
RTI International, Research Triangle Park, NC
Kimberly C. Peterson
RTI International
Mobile device use is increasingly common among young adults 18-29 (86% report owning a smartphone) (Anderson, 2015). As mobile use has increased so has mobile survey response. While research on the impact of mobile responding is mixed, there is some suggestion that it can lead to data quality issues such as increased break-offs (Mavletova, 2013). Data quality concerns may be compounded when surveys contain sensitive questions. Previous research found respondents trust mobile devices less with regard to data confidentiality, but saw mixed results when examining response to sensitive items by device type (i.e., mobile, PC) (Mavletova and Couper, 2013). However, this study focused on attitudes toward deviant behavior, alcohol consumption and income; findings may differ with other sensitive topics, such as sexual assault. In this paper we explore differences in item nonresponse and breakoff rates for sensitive questions related to sexual harassment and assault to better understand how survey estimates may differ by device type (i.e., mobile, non-mobile). We conclude with a discussion of findings and the implications for the design of sensitive web surveys of college students.