Web surveys generally attain lower response rates than other modes of data collection, and response rates seem to be declining over time. Web surveys are also at risk of other forms of nonresponse, such as breakoffs, which happen less frequently in other modes. These breakoffs commonly occur on informed consent screens, often required by IRBs, before they have a chance to get to the survey content. There are many methods used
(prenotification, incentives, etc.) to try to increase participation and reduce breakoffs. This paper focuses on two experiments designed to assess the impact in increasing participation and reducing breakoffs. These two experiments were implemented on the Sustainability Cultural Indicators Program (SCIP) annual survey at the University of Michigan. The SCIP program monitors the U-M’s progress in moving toward a culture of sustainability. Progress is determined by tracking a set of cultural indicators derived from responses to a yearly campus-wide sustainability questionnaire. The sample design includes a rotating panel design for undergraduate students and a yearly independent cross-section of faculty, staff and students. The first experiment examines if “celebrity endorsement” increases participation. As part of the final email reminder respondents were randomized to receive a regular reminder with a link to the survey or a reminder that also contained a link to a video of a head coach from the U-M Department of Athletics encouraging non-respondents to participate. The second experiment looks at the design of the informed consent screen. One group was presented a screen appearing as a traditional informed consent form. The other group was presented a screen with the most important items visible and the rest of the information available under a series of accordion menus.