Academic Data Analytics

Last updated: September 25, 2024

Appropriate Collection and Use of Academic Data Analytics

Note: This document will be revised as needed to address the strategic priorities of the University, evolving privacy requirements, and/or emerging topics such as automated decision-making, machine learning, etc.

1. Purpose

To establish clear requirements for the appropriate and ethical processing of academic data analytics, as described in this policy. Specifically, the use of academic data for analytics must align with these goals to:

  • Customize teaching and learning experiences to help students achieve their learning goals.
  • Match academic support and services with students who are eligible.
  • Align advising practices based on student interest and needs.
  • Improve academic planning tools to increase persistence and retention and reduce the time it takes to finish a degree.
  • Improve central strategic planning and related processes for enrollment predictions, financial aid, tuition planning, and degree-to-career pathways.

2. Scope and Applicability

All academic data at the UW regardless of how or where the data are collected or processed. All areas of the UW including, but not limited to UW Medicine, UW Tacoma, and UW Bothell.

3. Definitions

Academic data analytics

The collection, transformation, analysis, modeling, and use of academic data to improve learning, retention, and academic planning. This includes data that are created in the process of aggregating data elements. This may also include statistical modeling or machine learning techniques.

Personal Data

Refer to the Glossary of Privacy Terms.

4. Roles and Responsibilities

UW Personnel Who Create or Consume Academic data analytics

Personnel must ensure that:

  • Analytics are valid and effective, UW Privacy Principles and security policies are upheld, and governance structures are in place to ensure that such activities do not compromise UW’s values and policies.
  • Meaning is extracted from analytics in a manner that leads to positive outcomes and action with students.
  • Analytics are used effectively to support academic departments and units that strive to customize teaching and learning experiences.
  • Institutional planning and resources allocation for enrollment, financial aid, tuition planning, and degree-to-career pathways are supported.

Academic Data Domain Council

The Academic Data Domain Council is charged with developing a cohesive approach to stewarding the University’s academic data to derive the most value from academic data while managing risks.

Faculty Councils

Faculty councils serve as deliberative and advisory bodies for all matters of University policy as stated in the Faculty Code and Governance, Councils & Committees, Chapter 42.

Privacy Steering Committee

The Privacy Steering Committee is charged with helping uphold the UW values and guiding strategic decisions about the purpose and use of personal data. The committee will exercise oversight for the purpose and appropriate use standards for academic data analytics in consultation with applicable governance groups.

5. Types of Data Used for Academic Data Analytics

Data that are currently being used and permitted for ongoing use in academic data analytics (at an individual, de-identified, pseudonymized, and/or aggregate level) and for the specific goals outlined in Section 1 include:

  • Enrollment and application data such as degree program affiliation, academic probationary status, campus affiliation, and demographics provided by the student.
  • Data collected from clearinghouses, data warehouses shared with peer institutions, state agencies or third parties who a) have a formal or official relationship with the UW, and b) permit the use of data for academic data analytics.
  • Transcript data such as data from past and current courses, including grade data.
  • Teaching and learning tools data such as viewing patterns, number of discussion board posts, and logins.
  • Student academic information systems data such as EARS, MyUW, MyPlan, Canvas, Panopto and other student systems at the enterprise, college, school, or department level.
  • Financial Aid or economic status data solely for the purpose of awarding and management of aid, including but not limited to need-based aid and scholarships.
  • Card swipe data for documenting engagement and use of academic services, such as attendance to tutoring sessions.
  • Student and alumni career data such as career paths, earnings, and wages in connection with degrees.

Data that are not permitted for use in academic data analytics at an individual level include, but may not be limited to:

  • Data about an individual’s health, disabilities, immunizations or vaccinations, and use of student health services.
  • Formal complaints data made by students and personnel.
  • Affiliations data that are not directly related to academic success, such as religious or political affiliations.
  • Social media activity, such as data about student activity on third-party social networking sites.
  • Any records or information relating to criminal offenses, citizenship and/or immigration status, political opinions, religious or philosophical beliefs, trade union membership, genetic or biometric data used to identify a natural person, health, sex life, or sexual orientation.
  • All financial aid or economic status data and copies of the data must be limited to what is required to administer institutional aid, should not be maintained at an individually identifiable level longer than the UW-approved retention schedule of the department of record.
  • Location data (other than the specific use of card swipe data described previously) related to a student’s physical location at any point in time or movements from one location to another location, such as geolocation data.
  • Student spending habits or purchasing data, such as food and dining purchases.

6. Validity and Efficacy

Assessment and refinement of modeling, analysis and practices must be an ongoing process to ensure valid results and useful and effective service delivery.

  • Modeling and analysis of academic data must be free from biases, and practices for mitigating bias in the application of academic data analytics will be encouraged.
  • The accuracy of the models must be closely scrutinized on a periodic basis by UW data scientists to ensure they are meeting an acceptable level of accuracy and are free of bias.
  • Algorithms and other analytical processes performed on academic data must be available for review by University Personnel.
  • Results must be available to support collaborations with other institutions so long as review does not expose student data.
  • Errors in the data must be corrected in the systems from which the data is sourced (as much as possible and practical) rather than in the systems in which the data are consumed, analyzed, or displayed. Data subjects’ requests to correct data must follow the UW Data Subject Request Process. Processes, other than data subjects’ requests, to correct data must be reviewed and approved by the Academic Data Domain Council.

To learn more about existing academic data analytics visit the Data Science projects and reports provided by UW-IT.

For questions or support about validity and efficacy of academic data analytics contact the Academic Data Domain Council at datagov@uw.edu.

7. Privacy by Design

The processing (such as collection, storage, access, use, and analysis) of personal data for academic data analytics must:

  • Uphold UW Values and Privacy Principles.
  • Follow University Privacy and Data Protection Policies.
  • Limit disclosure of identifiable data to those that have a legitimate educational purpose for having access to the data to perform their official job duties at the UW.
  • Ensure reports with de-identified, pseudonymized, anonymized, or aggregate analytics are at the n>5 level, unless there is a specific (such as a legal reason) for de-identified, pseudonymized, anonymized, or aggregate analytics data to be at n>10, to help protect confidentiality and decrease the likelihood of individual identification.
  • Include a privacy assessment if the data processing is considered high risk.
  • Be consistent with the commitments that the UW made and communicated to individuals in the privacy notice provided to them prior to data being collected.
  • Be limited to the minimum data needed to achieve the specific business purpose.
  • Use the nomenclature that is consistent with the nomenclature used for internal and mandated federal reporting purposes and/or approved by the Academic Data Domain Council.
  • Anonymize data or minimize the volume of identifiable and de-identified personal data used in analytics whenever possible.
  • Include an internal data processing memorandum of understanding with internal parties at UW or an appropriate data processing agreement with external/third parties to UW as described on the UW Privacy Office website.

For questions or support about privacy and appropriate use contact the UW Privacy Office at uwprivacy@uw.edu.

8. Security

Personnel that manage or utilize information systems with academic data and analytics are required to follow the UW information security policies for safeguarding UW institutional information.

9. Authors

Authors: Tom Lewis, Henry Lyle, Ann Nagel, Matt Winslow, Erin Guthrie

Reviewed by:

  • Faculty Council on Information Technology and Cybersecurity
  • Academic Data Domain Council

Approved by the Privacy Steering Committee
Approved by the University Privacy Officer
Approved by Institutional Data Steward for Academic Data

Published 10/11/2022.
11/15/2023: Removed reference to Academic Experience Design and Delivery in section 6.