Timing Is Everything: Discreetly Discouraging Mobile Survey Response Through the Timing of E-mail Contacts

Abstract

The proportion of web survey responses submitted from mobile devices, such as smartphones, is increasing steadily. This trend is problematic because mobile responses are associated with increased breakoffs, item nonresponse, and other data quality issues. Careful web survey design can mitigate some of these concerns, but cannot eliminate them entirely. As a result, survey practitioners typically prefer that respondents not respond via mobile devices. Web surveys can be programmed to block mobile responses, but this approach is discouraged because of its potential to increase nonresponse (Buskirk & Andrus, 2012). Ideally, researchers need a way to discourage mobile response without impacting response rates. In this paper we evaluate a strategy for discretely discouraging mobile responding. The Campus Climate Survey Validation Study Pilot Test, sponsored by the Bureau of Justice Statistics and the Office on Violence Against Women, is a survey of college students at nine U.S. institutions of higher education. Over 23,000 respondents completed the survey among a random sample of approximately 50,000 students. Although schedules vary across students and institutions, we suspect college students are less likely to respond via mobile devices during certain times of the day than others. For example, in the early evening on a Monday, they may be more likely to be using a computer to complete assignments, and thus less likely to respond via mobile devices. Using data on the day and time a response was submitted as well as the day and time a respondent was last emailed a request to complete the survey, we identify the times that are most likely to result in nonmobile responses. Because web survey response typically spikes immediately after invitations and reminders are sent, the findings of our research can be used to carefully time email contacts in an attempt to discretely discourage mobile responding among a college student sample.