This is excerpted from a longer document that includes additional details about data governance and the purpose and background for the creation of the principles. Refer to The Appropriate Use of Data for Learning Analytics Guiding Principles for additional details. These guiding principles were created by the Learning Analytics Data Use Subcommittee, endorsed by the Learning Analytics Roadmap Committee on May 13, 2019, and were approved by the Data Governance Council on May 28, 2019.

In this document:

• Definition of Learning Analytics
• Guiding Principles / Values
• Scope: In Scope & Out of Scope
• Roles & Responsibilities

UW-Madison’s Definition of Learning Analytics (LA)

Learning analytics is the undertaking of activities that generate actionable data from the learning environment intended to improve student outcomes by informing structure, content, delivery, or support of the learning environment.

Guiding Principles / Values

Students are real and diverse individuals, and not just their data or information. These principles — beneficence, transparency, privacy and confidentiality, and minimization of adverse impacts — aim to uphold the dignity of students while ensuring learning analytics are used to improve educational outcomes.

Beneficence

• Use LA to benefit student learning and success, while potentially improving teaching.
• Ensure good evidence and pedagogical reasons are the backbone for data collection and interpretation.
• Engage students as active agents in designing LA interactions that will support them.

Minimization of Adverse Impacts

• Promote inclusivity and equity to mitigate bias.
• Ensure that data are not exploited for unexpected or unwanted uses or outcomes.
• Continually evaluate and monitor LA tools, practices, and interventions to assess differential impact and overall efficacy.
• LA should be only one data point for consideration when taking action; it should not be the only input.

Transparency

• Be transparent regarding the policies and procedures for collection, access, and use of student learning data for LA purposes.
• Ensure LA are well defined and visible to stakeholders, including methods used to initiate interventions.

Privacy and Confidentiality

• Follow existing campus privacy, data governance and security standards and practices, including those pertaining to institutional data access and stewardship.
• Share data/information only with those who are authorized.
• Use data in the most anonymous format possible that is still useful.

Scope: In Scope & Out of Scope

What is included in teaching and learning data use and learning analytics activities?

As an institution, we are required to collect and retain data as part of the student record. Data Governance Policy classifies LA data as “restricted” because the unauthorized disclosure of personal student information could cause potential harm to individuals affected or a significant level of risk to the university. The Learning Analytics Data Use Subcommittee establishes a clear delineation of what is in scope and what is out of scope for both LA data and LA activities, detailed in these tables.

Roles & Responsibilities

The LA guiding principles apply to everyone. Every stakeholder group has a responsibility to understand this document. Specific stakeholder roles and responsibilities are provided for students, instructors, advisors and leadership.

All stakeholders have the responsibility to:

• Refer to data governance, cybersecurity, and other university and UW System policies around data access, use, and retention.
• Be aware of opportunities to engage with LA
• Generate and use LA data ethically
• Make inquiries when LA activities are unclear or produce real or perceived negative impacts on students

Students have the responsibility to:

• Be aware of how LA data are being used for their benefit
• Respect the privacy of other students’ data
• Consider insights or interventions presented to them via LA activities
• Generate their LA data ethically

Instructors have the responsibility to:

• Understand FERPA and their responsibility regarding systems and data subject to it
• Understand the LA data/results and their implications before acting
• Consult with the appropriate supervisor or support person when needed
• Consider insights provided by LA and act in a sensitive manner appropriate to the instructional role
• Inform students about centrally-available or instructor-created resources and training to help students understand LA data/results
• Use de-identified data whenever possible, especially with longitudinal datasets
• Only use enterprise systems unless authorized by the institution
• Track instructional interventions triggered by LA

Advisors (Academic/Career Services and Student Affairs) have the responsibility to:

• Understand FERPA and their responsibility regarding systems and data subject to it
• Understand the LA data/results and their implications before acting
• Consult with the appropriate supervisor or support person when needed
• Inform students about centrally-available or instructor-created resources and training to help students understand LA data/result
• Consider insights provided by LA and act in a sensitive manner appropriate to the advisor role
• Use de-identified data whenever possible, especially with longitudinal datasets
• Track advising interventions triggered by LA, and ensure appropriate communication among academic, student affairs and career services advisors

Program and Institution Leadership have the responsibility to:

• Understand FERPA and their responsibility regarding systems and data subject to it
• Develop and support the technical infrastructure underpinning LA data use
• Develop, review, and strive to ensure compliance with student intervention policies, procedures, and standards
• Provide LA resources and training to help data users understand data/results
• Clearly communicate how students can engage with LA presented to them
• Safeguard data access
• Collect and incorporate LA data for reporting and accreditation purposes as required
• Use de-identified data whenever possible, especially with longitudinal datasets
• Practice non-maleficence (do no harm)