The use of smartphones and tablets of various sizes and operating systems has posed numerous challenges for researchers, impacting survey response and data quality. Importantly, though, their use also provides opportunities to capture nuanced data helpful in understanding their impact. A key element of such device use is the mobility of the respondent; one may be in transit or in a public location while responding to a Web survey previously only completed on a computer. Thus far, little data has been collected to assess the impact of mobility on data quality. With these aims in mind, we report the findings of an experiment assessing feasibility of collecting geolocation from survey respondents, as well as the quality of data received. In doing so, we discuss implications for survey implem entation and measurement error in an increasingly mobile world.
Within a series of short web surveys of college students fielded in 2013, we assessed feasibility by asking respondents whether they would be willing to let researchers collect geolocation data in the future. Building off positive results from this analysis, we then experimentally tested methods of informed consent for geolocation capture. Respondents were randomly assigned to receive no request for geolocation, a request via the survey consent form, or a request via a separate question focusing only on this data capture. We captured geolocation data at every page of the survey for those consenting.
We will discuss the results of the experiment, laying out a recommendation for obtaining consent; comparing those who consented to those who did not; and assessing the quality of the geolocation data obtained. We will also present analyses examining how such information may be used in a broader context-- exploring whether or not the mobility of the respondent may be a key predictor in evaluating survey data quality.