A Scenario-based Exploration of Expected Usefulness, Privacy Concerns, and Adoption Likelihood of Learning Analytics.

Learning analytics has become a robust research area in the last decade, as innovative analytic models of learning data have been created with the goal of enhancing teaching and learning. However, barriers to large scale adoption of such technologies in higher education still exist. In recent years, a strand of research has begun to investigate stakeholders' expectations of learning analytics, hoping to find ways to integrate the innovations into everyday teaching practices. For instance, studies have investigated instructors' ideas about how learning analytics might be helpful, as well as concerns about student data privacy. However, most studies have taken a general approach rather than considering instructors' day-to-day experiences. Using survey methods, we presented instructors with hypothetical scenarios of learning analytics in use across disciplines, class sizes, teaching activities, and types of student data. We asked for ratings of both usefulness and privacy concerns for each proposed teaching situation. Our respondents considered scenarios involving learning outcomes-related data (e.g. grades) to be more useful than those that involve student interactions (e.g. language, social activity). In contrast, privacy concerns were lower for outcomes-oriented scenarios than interactions-focused scenarios. An interesting new finding was a negative correlation of usefulness and privacy; we discuss this in the context of instructors' possible cost-benefit reasoning. We reflect on our findings with respect to future efforts in developing and fielding learning analytics tools.