As a researcher who focuses on how US higher education institutions construct data systems to support students’ success, I find myself having the same conversation with campus stakeholders over and over. That conversation involves two key questions: what do we do with all this data and do we actually need it? Indiscriminate data capture and indefinite data storage––with an eye on some obscure future application––bring costs that campus stakeholders have yet to wrestle with:
- Financial costs such as building new networks, licensing or purchasing server space, and hiring and supporting data analysts;
- Environmental costs such as the energy required and the pollution produced by campus servers (as of 2015, estimates suggested that the information and communications technology ecosystem contributed to about 2 percent of global emissions); and
- Individual costs such as potential privacy violations and the removal of individuals’ autonomy as part of academic planning (for instructors) and study strategy development (for students)
If we use data resources to guide our governance and decision making, we also need to account for how we will be governed by data. What political, economic, and environmental consequences does data encourage us to accept through the velocity of its accumulation? Below, I offer three contemporary examples from my research and the questions they pose for institutional leaders about student rights and institutional futures. Debating and resolving these questions now can ensure that institutional leaders protect individuals’ privacy on campus, ensure individual autonomy, and reduce financial and environmental costs.
Autonomy, Privacy, and Classroom Governance
Surveillance technologies present concerns about privacy and individual autonomy for students, faculty, and staff. In research on faculty’s use of teaching dashboards, faculty members consistently expressed concern about how the data produced by the dashboard would be used to evaluate their teaching. Although evaluating teaching had not been the dashboard’s original purpose (which was to capture information about students’ behavioral engagement in the classroom), department chairs realized that the dashboard provided a window into the classroom. They could track attendance, view how often instructors used active learning strategies, and gain insight into how instructors reacted to the systems’ feedback.
This classroom intervention was unprecedented, and instructors insisted that it violated their rights to determine course curriculum and pedagogy. More important, it subjected them to scrutiny in new and invasive ways, and it potentially changed their teaching decisions during the course.
Pre-tenure and adjunct faculty reported increasing their use of the learning management system and implementing new external enforcement contingencies (such as requiring attendance as part of students’ grades) because they could see how both strategies would improve their metrics. This sort of increase, in turn, further entangles students into the digital ecosystem of instructional technologies. In one course, when the instructor created a new attendance requirement after previously having posted lecture videos, students had to make quick changes to their work schedules and child care arrangements to ensure they were present for the lecture.
These findings invite the following question for institutional leaders: what expectation of individual autonomy and privacy should instructors have regarding pedagogical and curricular decision-making?
Indefinite Storage and Ecological Impact
After institutions capture and store data from the classroom, it is unclear for how long the institutions hold on to trace logs. In nearly all of the policies that I reviewed with Dr. Carrie Klein to understand how institutions construct students’ right to data privacy, we observed no clear timelines for when institutions would delete or remove information. Campus information technologists have expressed concern about how much data the campus realistically expects to track; some states have gone so far as to construct cradle-to-grave longitudinal data systems, of which campus information is one small component of the data avalanche.
Education researchers, like all social scientists engaged in statistical analyses, have trouble separating signal from noise in data. Contemporary debates in the field of learning analytics and big data in education reflect longstanding debates in education research about how to measure learning, change, and development. Can we observe, for example, something important about an individual’s meta-cognition from the trace data they leave in an educational tool? The best answer, at the moment, seems to be “maybe” under certain conditions in which we can control a good deal of the context and activity. As researchers try to refine their ability to identify learning, the response of institutions has been to collect as much data as possible and to retain it indefinitely.
These facts invite institutional leaders to consider the following question: how long should we expect to keep student behavioral data given the cost of data storage and individuals’ right to control their representations?
Institutional Trust and Opting Out of the Data Deluge
In my most recent research project, I engage with instructors who use similarity detectors in their classrooms, to better understand how they learn about data management and what they communicate to their students about the technology. Consistently, participants express frustration about the ways in which the technology erodes trust––between students and the instructor and between individuals and the institution. Many institutions now require that instructors include a syllabus statement for similarity/plagiarism checkers, detailing which data will be stored where, by whom, and what control students have over their writing.
That institutions recognize that students should have control over their data when it is handed off to a third party (effectively for monetization) suggests that institutions can create governance technologies (such as opt-out mechanisms for Turnitin) that center individual agency. That institutions have not yet extended these same governance technologies to their own activities raises important questions about privacy and autonomy. Recent research suggests that students generally do not understand their rights and entitlements to ensure their data privacy and that women and students of color are more distrustful of how institutions might use their data.
These considerations indicate the following question for institutional leaders: can institutions build trust about data use without empowering individuals to control the extraction, storage, and analysis of their data traces?
Institutions’ concern about opt-out mechanisms as one way to allow individuals to control their data governance stems from this solution’s potential to create systematically missing data. But we should also acknowledge how this approach (and similar approaches to allowing students, faculty, and staff to control their data representations) allows individuals to address which data is stored, for how long, and to what end. By empowering individuals to manage their own data, we allow them to make their own calculations about the economic, political, and ecological consequences of their data assets.
