Author: Kyle M. L. Jones (MLIS, PhD) • Indiana University-Indianapolis (IUPUI)

Executive Summary

Some higher education institutions (HEIs) are developing their educational data mining and analytic capacities, while others are purchasing technologies to support their interests in analytics. These efforts involve a broad array of practices that many consider to be a form of learning analytics. Learning analytics are designed to improve student learning outcomes, provide students with just-in-time resources, and make predictions to direct student behaviors toward actions that lead to academic success. Institutions also use learning analytics to discover cost savings and to increase efficiencies. To support learning analytics initiatives, institutions are increasing both their collection of students’ data and information and the types of student data they collect.

Learning analytics raises significant privacy issues, such as the potential for bias, racial and other discrimination, and reductive analyses that could compromise or even foreclose students’ future education and job opportunities. The increasing collection of student data raises red flags about whether such practices invade privacy, but arguably more notable is the power that institutions gain over their students. Universities can use their ubiquitous data and technology infrastructures to capture student life and behaviors at a granular level, through techniques such as mandated digital learning systems and tracking students’ locations via radio-frequency identification chips in their university ID cards. Institutional rules, norms, and expectations—all formally codified or verbally relayed—hold sway over students. The power differential between students and institutional actors, such as administrators, faculty, and staff, is immense. With learning analytics, institutions use data visualization, predictive measures, and targeted messaging to mold student behaviors in ways that potentially and negatively affect their autonomy and direct their choices. Students are mostly unaware of these techniques and cannot remove themselves from a university’s network of technology. The result is that students have little agency to contest learning analytics driven by privacy- invading data practices and technologies.

Policymakers and higher education stakeholders could benefit from deeper understanding of student privacy, but there is scant literature explaining its value, and arguments rely too heavily on compliance with federal law. To foster this understanding, this brief explains that student privacy is rooted in contextual values and expectations, is critical to intellectual freedom, and supports students in their various institutional relationships. If learning analytics is to mature in alignment with the privacy protections and ethical practices that students and other stakeholders expect, HEIs must commit to the following actions:

• Higher education actors should review analytic initiatives for fairness, bias, and privacy problems.

• The academic community must reflect on the consequences of privacy-invading technologies, including their ability to prop up administrative interests.

• Stakeholders, including students, need to participate in the co-design of learning analytics to ensure an equitable, agreeable vision and implementation of the technology.