Addressing the wicked problem of learning data privacy though principle and practice
Download as PPTX, PDF1 like2,053 views
The document outlines the principles and practices surrounding learning data privacy at the University of California, emphasizing ownership, ethical use, transparency, and security of student data. It highlights the importance of learning analytics in improving student outcomes and institutional decision-making while advocating for faculty and students' rights over their data. The document also underscores the need for interoperability and collaboration between institutions and service providers to ensure effective use of learning data.
![We Are Not Alone...
● [2011] - Asilomar Convention for Learning Research in Higher Education - http://asilomar-highered.info/
● [2012] - Asilomar II: Student Data and Records in the Digital Era - https://sites.stanford.edu/asilomar/
● Responsible Use of Student Data in Higher Education - http://ru.stanford.edu/ (Stanford CAROL & Ithaka
S+R)
● DELICATE Framework - http://www.laceproject.eu/blog/ethics-privacy-in-learning-analytics-a-delicate-issue/
(Learning Analytics Community Exchange)
● IMS Global Learning Data & Analytics Key Principles - http://www.imsglobal.org/learning-data-analytics-key-
principles](https://image.slidesharecdn.com/1400cnistringerroom11formatted16x9-180703111052/85/Addressing-the-wicked-problem-of-learning-data-privacy-though-principle-and-practice-23-320.jpg)
Addressing the wicked problem of learning data privacy though principle and practice
- 2. The Wicked Problem of Learning Data Privacy Jenn Stringer, MLIS University of California, Berkeley IMS Global: Samantha Birk, IMS Global, John Fritz, University of Maryland, Baltimore County, Oliver Heyer, University of California, Berkeley, Virginia Lacefield, University of Kentucky, Virginia Lacefield, University of Kentucky, Adam Recktenwald, University of Kentucky, Marianne Schroeder, University of British Columbia University of California (UC): Mary Ellen Kreher, UC Office of the President, Jim Phillips UC, Santa Cruz, James Williamson, UC, Los Angeles
- 3. Academic Data Personally identifiable records, e.g, transcripts (course work, GPA, major), enrollments, academic plan, SAT scores etc. Student Information & Advising Systems Learning Data Personally identifiable user activity, e.g., Page views, Discussion posts, Quiz responses, Video views etc. recorded in LMS’s and other 3rd party learning applications Learning Record Store Institutional Data Aggregate, often deidentified, historical records e.g. Graduation rates, yield, application data, demographics, race/ethnicity Enterprise Data Warehouse Learning Data in Context 3 “Learning data refers to data generated by students, faculty, and/or staff that relates to and documents the teaching and learning experience and academic achievement. It can be used alone or combined with the student record and other data points to support student success and research.” 1 1 https://www.imsglobal.org/learning-data-analytics-key-principles
- 4. Learning Analytics Defined “Learning analytics is the measurement, collection, analysis and reporting of data about learners and their contexts, for purposes of understanding and optimising learning and the environments in which it occurs.”2 21st International Conference of Learning Analytics & Knowledge, Banff, Alberta 2011
- 5. ● increase ability to make institutional decisions ● impact student outcomes ● empower students to make changes to their behavior that positively affects their learning ● enable faculty to support students and make changes to their courses based on data ● support faculty teaching and pedagogy ● support educational research Why do we care?
- 6. Learning Data ● “Old Days” ○ local hosting meant local logs ○ ad-hoc reporting mainly for systems issues ● “Cloud SaaS” ○ logs not local and not accessible ○ vendors use to improve systems and troubleshoot issues ● Contracts ○ if we have them -- not always specific about ownership of this kind of data ○ they do have security and privacy language ● No Contract? ○ goodluck! - at the whim of the third party provider no security/privacy guarantees
- 7. ● What: “ Clickstream Data” “Logfile Data” ● Who: Vendors, the institution, libraries, publishers, other third-parties 7 Someone is Collecting A LOT of data
- 8. 8
- 9. 9
- 10. “If it is free, you are not the customer. You are the product.”4 4http://blogs.harvard.edu/futureoftheinternet/2012/03/21/meme-patrol-when-something-online-is-free-youre-not-the-customer-youre-the-product/
- 12. 12
- 13. 13
- 14. A little better …
- 15. 15
- 16. A little better …
- 17. Learning Data ● Just the data please ○ even if we have “ownership” we need “access” ○ logfile ‘dumps” for historic reporting (nightly, monthly, etc.) ○ realtime for early warning systems, advising, student alerts, etc. ● Standards ○ LTI - enables interoperability ○ Caliper and xAPI - defines learner activity to enable analysis across systems
- 18. The University of California should have a say in how suppliers collect, use, and manage our data.
- 19. ● 10 Campuses ● 150 Academic Disciplines ● 238,700 Students ● 44,517 Academic Staff Who are “we”? The UC System
- 20. University of California: Learning Data Privacy Principles 1. Ownership: The University of California (UC), its faculty, and students retain ownership of the data and subsequent computational transformations of the data they produce. Individual data owners have the right to determine how their data will be used. The UC acts as stewards of data on behalf of its faculty and students. 2. Ethical Use: Learning data collection, use, and computational transformation are governed by pedagogical and instructional concerns, with an aim toward student success through prescriptive, descriptive, or predictive methodologies. As with grades and other sensitive data, uses of learning analytics should be pursued on a “need to know” basis. 3. Transparency: Data owners have a right to understand the specific methods and purposes for which their data are collected, used and transformed, including what data are being transmitted to third-party service providers (and their affiliated partners) and the details of how algorithms are applied that shape summaries, particularly outputs and visualizations. 4. Freedom of Expression: Faculty and students retain the right to communicate and engage with each other in the learning process without the concern that their data will be mined for unintended or unknown purposes. 5. Protection: Stewards, on behalf of data owners, will ensure learning data are secure and protected in alignment with all federal, state, and university regulations regarding secure disposition. 6. Access and Control: Data owners have the right to access their data. Given that faculty and students own their learning data and share in its disposition, access to and ultimate authority and control of the data rests with the faculty and student owners, and the data stewards acting on their behalf. Data retention access and control practices will be governed under UC policies and supplier contractual agreements. 20 http://bit.ly/UCLearningDataPrivacyPrinciples
- 21. University of California:Learning Data Recommended Practices 1. Ownership: Service providers will recognize learning data ownership and access as a right of the faculty and students. 2. Usage Right: Through a user’s profile setting, service providers will enable users to control the use of their intellectual property. Thus, it will be the user’s choice to grant terms such as, “a royalty-free, transferable, perpetual, irrevocable, non-exclusive, worldwide license to reproduce, modify, publish, publicly display, make derivative works.” 3. Opt-in: Other than those data elements distinctly required for instruction, where appropriate, students will have a choice about the use of learning data collected by faculty and service providers in an "opt in" rather than "opt out" approach. 4. Interoperable Data: Service providers will provide learning data to the institution in recognized standard interoperability format(s) to minimize integration costs, support cross-platform and cross-application uses, and promote institutional and academic analysis and research. 5. Data without Fees: Service providers will not charge the faculty, students, or other university learning data stewards for the right of access, including the delivery of these data to the University. 6. Transparency: Service providers will inform the UC about the learning data they collect and how these data will be used, which in the course of an academic term shall be based on pedagogical concerns and curricular improvement. 7. Service Provider Security: All service provider platforms on which student learning data are stored will conform with UC and state mandated security procedures governing the reporting of unexpected incidents and corrections that may occur. 8. Campus Security: UC learning data stewards will ensure that all faculty and student data are stored securely in conformance with University data security policy. Learning data stewards will report any learning data security incidents as appropriate to faculty and students, and will provide information about their remedy. 21
- 22. IMS Global Learning Data & Analytics Key Principles 1. Ownership: Faculty, staff, and students generate and own their learning data. As governed by institutional policies, individuals, being owners of the data they generate, have the right to access, port, and control the disposition of their data stored by the institution, its service providers, and their affiliated partners. 2. Stewardship: As stewards of learning data, institutions should have a data governance plan and governance policies that protect the data and the interests of its owners. These should transcend, but encompass, existing protocols, such as IRB. 3. Governance: Learning data use and retention will be governed by institutional policies, and faculty and students retain the right of data access and retrieval. 4. Access: Learning data, whether generated locally or in a vendor-supplied system, is strategic to an institution’s business and mission and must be available to the institution. 5. Interoperability: The collection, use, and access to learning data requires institutional and supplier collaboration, which is dependent upon interoperability standards, protocols, data formats, and content to achieve institutions goals. 6. Efficacy: Learning data collection, use, and computational transformation is aimed at student and instructor success and instructional concerns through prescriptive, descriptive, or predictive methodologies. 7. Security & Privacy: Individuals’ security and privacy relating to collecting, using, and algorithmically transforming learning data is fundamental and must not be treated as optional. It must also be balanced with the effective use of the data. 8. Transparency: Individuals have the right to understand the specific reasons, methods, and purposes for which their learning data is collected, used, and transformed. This includes any learning data being shared with third-party service providers and other institutional affiliates or partners. Individuals also have the right to know how their data is transformed and/or used thru processes such as summative or algorithmic modifications, particular outputs, and visualizations. 22 https://www.imsglobal.org/learning-data-analytics-ke principles
- 23. We Are Not Alone... ● [2011] - Asilomar Convention for Learning Research in Higher Education - http://asilomar-highered.info/ ● [2012] - Asilomar II: Student Data and Records in the Digital Era - https://sites.stanford.edu/asilomar/ ● Responsible Use of Student Data in Higher Education - http://ru.stanford.edu/ (Stanford CAROL & Ithaka S+R) ● DELICATE Framework - http://www.laceproject.eu/blog/ethics-privacy-in-learning-analytics-a-delicate-issue/ (Learning Analytics Community Exchange) ● IMS Global Learning Data & Analytics Key Principles - http://www.imsglobal.org/learning-data-analytics-key- principles
- 24. IMS Global Learning Data & Analytics Key Principles https://www.imsglobal.org/learning-data-analytics-key-principles 1. Ownership: Faculty, staff, and students generate and own their learning data. As governed by institutional policies, individuals, being owners of the data they generate, have the right to access, port, and control the disposition of their data stored by the institution, its service providers, and their affiliated partners. 2. Stewardship: As stewards of learning data, institutions should have a data governance plan and governance policies that protect the data and the interests of its owners. These should transcend, but encompass, existing protocols, such as IRB. 3. Governance: Learning data use and retention will be governed by institutional policies, and faculty and students retain the right of data access and retrieval. 4. Access: Learning data, whether generated locally or in a vendor-supplied system, is strategic to an institution’s business and mission and must be available to the institution. 5. Interoperability: The collection, use, and access to learning data requires institutional and supplier collaboration, which is dependent upon interoperability standards, protocols, data formats, and content to achieve institutions goals. 6. Efficacy: Learning data collection, use, and computational transformation is aimed at student and instructor success and instructional concerns through prescriptive, descriptive, or predictive methodologies. 7. Security & Privacy: Individuals’ security and privacy relating to collecting, using, and algorithmically transforming learning data is fundamental and must not be treated as optional. It must also be balanced with the effective use of the data. 8. Transparency: Individuals have the right to understand the specific reasons, methods, and purposes for which their learning data is collected, used, and transformed. This includes any learning data being shared with third-party service providers and other institutional affiliates or partners. Individuals also have the right to know how their data is transformed and/or used thru processes such as summative or algorithmic modifications, particular outputs, and visualizations.
- 25. IMS Global Learning Data & Analytics Key Principles https://www.imsglobal.org/learning-data-analytics-key-principles 1. Ownership: Faculty, staff, and students generate and own their learning data. As governed by institutional policies, individuals, being owners of the data they generate, have the right to access, port, and control the disposition of their data stored by the institution, its service providers, and their affiliated partners. 2. Stewardship: As stewards of learning data, institutions should have a data governance plan and governance policies that protect the data and the interests of its owners. These should transcend, but encompass, existing protocols, such as IRB. 3. Governance: Learning data use and retention will be governed by institutional policies, and faculty and students retain the right of data access and retrieval. 4. Access: Learning data, whether generated locally or in a vendor-supplied system, is strategic to an institution’s business and mission and must be available to the institution. 5. Interoperability: The collection, use, and access to learning data requires institutional and supplier collaboration, which is dependent upon interoperability standards, protocols, data formats, and content to achieve institutions goals. 6. Efficacy: Learning data collection, use, and computational transformation is aimed at student and instructor success and instructional concerns through prescriptive, descriptive, or predictive methodologies. 7. Security & Privacy: Individuals’ security and privacy relating to collecting, using, and algorithmically transforming learning data is fundamental and must not be treated as optional. It must also be balanced with the effective use of the data. 8. Transparency: Individuals have the right to understand the specific reasons, methods, and purposes for which their learning data is collected, used, and transformed. This includes any learning data being shared with third-party service providers and other institutional affiliates or partners. Individuals also have the right to know how their data is transformed and/or used thru processes such as summative or algorithmic modifications, particular outputs, and visualizations.
- 26. Libraries and Learning Data/Analytics “Though few academic libraries are encountering it just yet, it is only a matter of time before higher education institutions integrate learning analytics at every level of the organization.”8 “Learning analytics initiatives pose a myriad of ethical questions. For example, are institutions who possess learning data required to act on it? Might learning data be used to “profile” students?”8 “At institutions that have committed to a learning analytics future, librarians can also ask questions to clarify the library’s role as well as advocate for library inclusion in learning analytics processes.”8 8 Bell, Steven. Keeping up with … Learning Analytics. ACRL Blog. http://www.ala.org/acrl/publications/keeping_up_with/learning_analytics 9 Oakleaf, Megan. “Getting Ready and Getting Started: Academic Librarian Involvement in Institutional Learning Analytics Initiatives.” Journal of Academic Librarianship. 42(4). 2016.
- 27. jisc.ac.uk One Castlepark, Tower Hill, Bristol BS2 0JA [email protected] T 020 3697 5800 Jenn Stringer, MLIS University of California, Berkeley