The field of learning analytics emerged in the early 2010s as researchers and practitioners sought to make use of the large amounts of digital trace data being produced by students in online and technology-enhanced learning environments. Early work in learning analytics focused on using basic LMS and student information system data to identify patterns between student behaviors and outcomes. However, critics argued this work was limited and did not address deeper learning challenges. By the mid-2010s, the field was establishing itself as a distinct "tribe" with its own conferences and journals, though it drew scholars from various disciplines and lacked a unified theoretical framework. Most recent reviews find the field is moving from predictive modeling toward a deeper understanding of the student experience but that evidence of impact on
1. Here be dragons: mapping the
(un)chartered in learning analytics
Virtual Presentation, School of Educational Studies and Leadership (EDSL),
(Te) Kura Mātauranga me te Rangatiratanga,
Tuesday, 7 August, 2018 Paul Prinsloo
University of South Africa (Unisa)
@14prinspImage credit: https://commons.wikimedia.org/wiki/File:Lenox_Globe_(2)_Britannica.png
3. Acknowledgement
I do not own the copyright of any of the images in this
presentation. I therefore acknowledge the original copyright
and licensing regime of every image used.
This presentation (excluding the images) is licensed
under a Creative Commons Attribution 4.0 International
License.
4. Source credit: https://www.theatlantic.com/technology/archive/2013/12/no-old-maps-actually-say-here-
be-dragons/282267/
But a globe does.
That’s right: One globe—just one—contains the words Hic
sunt dracones. Called the Hunt-Lenox Globe, it was built in
1510, making it one of the first European globes ever made
Not a single old paper map presents those exact words—
“Here be dragons”— in the margins or otherwise. Nor does
any paper map include “Hic sunt dracones,” the words’ Latin
equivalent.
6. Overview of presentation
• In the beginning… The emergence of the tribe
• The state of the tribe/field in 2014
• Fast forward to 2018
• And ethics?
• Mapping (some of) the unchartered
1. How is the field/context changing
2. The role of automated-decision making systems in learning analytics
3. The role of regulation/law/policy
4. Is learning analytics research or Research and what about the ethics?
5. Understanding the complexity of student success and how learning
analytics fit into describing, understanding, predicting and prescribing
6. Student-centered learning analytics
• (In)conclusions
8. Source credit: https://tekri.athabascau.ca/analytics/
“Learning institutions and corporations make little use of the
data learners ‘throw off’ in the process of accessing learning
materials, interacting with educators and peers, and creating
new content. In an age where educational institutions are under
growing pressure to reduce costs and increase efficiency,
analytics promises to be an important lens through which to
view and plan for change at course and institution levels”
(emphasis added).
9. Source credit: https://tekri.athabascau.ca/analytics/
“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.”
12. • Who are the ‘gate-keepers’, the voices in
learning analytics? How does it matter?
• What are the ‘rules”, ‘tribal culture’?
• How does the Annual conference, the
acceptance rate, the proceedings shape the
field?
• How does the Journal of Learning Analytics
shape the (future of) the field?
Let us consider, for a moment…
• Where are the centers of learning analytics located
(geopolitically) and institutionally and how does this matter?
• What/who are the competing fields/tribes and how does
learning analytics relate with them?
13. Dawson, S., Gašević, D., Siemens, G., & Joksimovic, S. (2014, March). Current state and future trends: A
citation network analysis of the learning analytics field. In Proceedings of the fourth international
conference on learning analytics and knowledge (pp. 231-240). ACM.
2014
14. Dawson, S., Gašević, D., Siemens, G., & Joksimovic, S. (2014, March). Current state and future trends: A
citation network analysis of the learning analytics field. In Proceedings of the fourth international
conference on learning analytics and knowledge (pp. 231-240). ACM.
Learning analytics as “field draws on assorted
theory and methodologies from disciplines as
diverse as education, psychology, philosophy,
sociology, linguistics, learning sciences,
statistics, machine learning/artificial intelligence
and computer science” (p. 232)
Describing the tribe
15. Dawson, S., Gašević, D., Siemens, G., & Joksimovic, S. (2014, March). Current state and future trends: A
citation network analysis of the learning analytics field. In Proceedings of the fourth international
conference on learning analytics and knowledge (pp. 231-240). ACM.
Network of all authors in the LAK conferences coded by disciplinary
background. Red: Computer Science; Blue: Education; Green: Other
(Industry, Engineering; Linguistics; or Business) (nodes sized by degree
centrality) (2011-2013)
• Computer Science – 51%
• Education – 40%
• “Missing” - machine learning,
artificial intelligence, statistics, and
data mining
• Relatively few inter-disciplinary nodes
16. Dawson, S., Gašević, D., Siemens, G., & Joksimovic, S. (2014, March). Current state and future trends: A
citation network analysis of the learning analytics field. In Proceedings of the fourth international
conference on learning analytics and knowledge (pp. 231-240). ACM.
“While LMS and SIS data can provide insight into how to improve
teaching and learning, this level of focus is not suitably aligned
with the substantial challenges that face all levels of education –
many of which require a systemic and integrated response. For
example, while it is helpful to note that students who regularly log
into a LMS may perform better than their less active peers, this
information is not suitable for developing a focused response to
poor performing students. It is neither helpful nor productive to
simply tell under-performing students to log in more
frequently” (p. 231; emphasis added)
What the tribe does: Early learning analytics
17.
18. Dawson, S., Gašević, D., Siemens, G., & Joksimovic, S. (2014, March). Current state and future trends: A
citation network analysis of the learning analytics field. In Proceedings of the fourth international
conference on learning analytics and knowledge (pp. 231-240). ACM.
“Learning analytics to date has served to
identify a condition, but has not advanced to
deal with the learning challenges in a more
nuanced and integrated manner” (p. 232)
What the tribe should do…
19. Dawson, S., Gašević, D., Siemens, G., & Joksimovic, S. (2014, March). Current state and future trends: A
citation network analysis of the learning analytics field. In Proceedings of the fourth international
conference on learning analytics and knowledge (pp. 231-240). ACM.
Cross-tabulation of the authors’ home disciplines (i.e., numbers
in the circles represent the number of authors) with both
research methods reported in the journal papers and types of
journal papers (2011-2013)
20. Dawson, S., Gašević, D., Siemens, G., & Joksimovic, S. (2014, March). Current state and future trends: A
citation network analysis of the learning analytics field. In Proceedings of the fourth international
conference on learning analytics and knowledge (pp. 231-240). ACM.
Cross-tabulation of the authors’ home disciplines (i.e., numbers
in the circles represent the number of authors) with both
research methods reported in the conference papers and types
of conference papers (2011-2013)
21. Viberg, O., Hatakka, M., Bälter, O., & Mavroudi, A. (2018). The Current Landscape of Learning Analytics in
Higher Education. Computers in Human Behavior.
2018
22. Viberg, O., Hatakka, M., Bälter, O., & Mavroudi, A. (2018). The Current Landscape of Learning Analytics in Higher
Education. Computers in Human Behavior.
Educational Data-Mining Learning Analytics
Automated discovery Focus on leveraging human
judgement
Automated adoption Developed to inform
instructors and learners
Reductionist frameworks –
“they reduce phenomena to
components and focus on the
analysis of individual
components and relationships
between them” (p. 1)
“Stronger focus on
understanding complex
systems as wholes”
23. Viberg, O., Hatakka, M., Bälter, O., & Mavroudi, A. (2018). The Current Landscape of Learning Analytics in Higher
Education. Computers in Human Behavior.
Four propositions of learning analytics
(Ferguson & Clow, 2017)
1)improve learning outcomes
2)support learning and teaching?
3) deployed widely; and
4) used ethically?
To what extent does learning analytics
To what extent is learning analytics
24. Viberg, O., Hatakka, M., Bälter, O., & Mavroudi, A. (2018). The Current Landscape of Learning Analytics in Higher
Education. Computers in Human Behavior.
57%
“There are no dominating theories;
rather there is a plethora of theories
used to explain different aspects of LA.
[…] While the theoretical development
of the field is still in its infancy, a few
field-specific theories have been
developed and applied” (p. 7)
25. Viberg, O., Hatakka, M., Bälter, O., & Mavroudi, A. (2018). The Current Landscape of Learning Analytics in Higher
Education. Computers in Human Behavior.
“Our results also demonstrate
that the overall potential of LA
is so far higher than the actual
evidence” (p. 10)
26. Viberg, O., Hatakka, M., Bälter, O., & Mavroudi, A. (2018). The Current Landscape of Learning Analytics in Higher
Education. Computers in Human Behavior.
Learning analytics evidence across the years 2012–
2018 (%).
27. Viberg, O., Hatakka, M., Bälter, O., & Mavroudi, A. (2018). The Current Landscape of Learning Analytics in Higher
Education. Computers in Human Behavior.
• “only 19% of the reviewed studies used both qualitative
and quantitative methods of data analysis, with an increase
in 2017”
• “Predictive methods have been one of the dominating
methods for several years. However, since 2016 the use of
these methods has considerably decreased. The decrease,
together with an increase of relationship mining methods
and the rather stable use of methods for the distillation of
data for human judgement, suggests that LA research in HE
is shifting from prediction of, e.g., retention and grades,
towards a deeper understanding of students’ learning
experiences.”
28. Viberg, O., Hatakka, M., Bälter, O., & Mavroudi, A. (2018). The Current Landscape of Learning Analytics in Higher
Education. Computers in Human Behavior.
• “so far there is little evidence (9%) that the research
findings demonstrate improvements in learning
outcomes, including knowledge acquisition, skill
development and cognitive gains, as well as learning
support and teaching”
29. Viberg, O., Hatakka, M., Bälter, O., & Mavroudi, A. (2018). The Current Landscape of Learning Analytics in Higher
Education. Computers in Human Behavior.
“It is worrying that more than 80% of the papers do
not mention ethics at all. Moreover, there are only
few studies that approach ethical issues (e.g., data
privacy and security, informed consent) in a
systematic way. However, we should not jump to the
conclusion that most studies are done in an unethical
way, but we call for more explicit reflection on ethics
to rise in the coming years. The increase of the
studies that reflect on the ethical issues for the year
2017 (36%) might indicate that there is already a
positive move in this direction” (emphasis added).
30. Prinsloo, P., & Slade, S. (2017). Ethics and Learning Analytics: Charting the (Un)Charted. In: Lang, Charles;
Siemens, George; Wise, Alyssa and Gašević, Dragan eds. Handbook of Learning Analytics. SOLAR, pp. 49–57.
2017
31. Prinsloo, P., & Slade, S. (2017). Ethics and Learning Analytics: Charting the (Un)Charted. In: Lang, Charles;
Siemens, George; Wise, Alyssa and Gašević, Dragan eds. Handbook of Learning Analytics. SOLAR, pp. 49–57.
“In a context where much is to be said for the potential
economic benefits (for both students and the institution)
of more successful learning experiences resulting from
increased data harvesting, we should not ignore the
possibilities of ‘data-proxy-induced hardship... when the
detail obtained from the data-proxy comes to
disadvantage its embodied referent in some way’(Smith,
2016, p. 16; also see Ruggiero, 2016; Strauss, 2016b; and
Watters, 2016)” (Prinsloo & Slade, 2017, p. 50; emphasis
added)
32. Prinsloo, P., & Slade, S. (2017). Ethics and Learning Analytics: Charting the (Un)Charted. In: Lang, Charles;
Siemens, George; Wise, Alyssa and Gašević, Dragan eds. Handbook of Learning Analytics. SOLAR, pp. 49–57.
In a recent overview of learning analytics practices in
the Australian context, Dawson, Gašević, and Rogers
(2016) report that the “relative silence afforded to
ethics across the studies is significant” (p. 3) and that
this “does not reflect the seriousness with which the
sector should consider these issues” (p. 33). The report
suggests that “It is likely that the higher education
sector has not been ready for such a conversation
previously, although it is argued that as institutions are
maturing, ethical considerations take on a heightened
salience” (p. 33).
33. Prinsloo, P., & Slade, S. (2017). Ethics and Learning Analytics: Charting the (Un)Charted. In: Lang, Charles;
Siemens, George; Wise, Alyssa and Gašević, Dragan eds. Handbook of Learning Analytics. SOLAR, pp. 49–57.
Interesting/important issues
• Educational triage – allocating (scarce) resources to where it
will make the most difference and ensure a return-on-
investment
• The responsibility of knowing and the moral duty to act once
we know
• Differentiating between essential information and ‘handy’
information – the dangers of collecting/scraping data that,
out of the original context, lose validity
• The mandate of higher education institutions and the moral
obligation to use student data
• The nuances of opting out
• Ethical issues that sit “outside of the law”
35. 1. How is the field/context changing
2. The role of automated-decision making systems
in learning analytics
3. The role of regulation/law/policy
4. Is learning analytics research or Research and
why/how does it matter?
5. Understanding the complexity of student success
and how learning analytics fit into describing,
understanding, predicting and prescribing
6. Student-centered learning analytics
36. Higher education has always collected, analysed
and used student data – so what has changed?
Image credit: https://en.wikipedia.org/wiki/Scholasticism
37. IN THE PAST AT PRESENT
Data sources Demographic and learning data at
specific points in the learning
journey: data application,
registration, class registers,
assignments, summative
assessment, personal
communication
Continuous, directed, gifted
and automated collection of
data from a range of data
sources – student
administration, learning
management system (LMS),
sources outside of the LMS
Data use Reporting purposes, operational
planning on cohort, group level by
management, institutional
researchers
Descriptive, diagnostic, predictive
and prescriptive on group/cohort
level
Plus individualised, often real-
time use of data to inform
pedagogy, curriculum,
assessment, student support by
faculty, students and support
staff
Who used the
data
(officially)?
Management, institutional
researchers, planners, quality
assurance and HR departments
Plus researchers, faculty,
students and support staff
1
38. IN THE PAST AT PRESENT
Who did the
collection,
analysis and
who used the
data
Humans Increasingly humans in
combination with algorithmic
decision-making processes
Temporal aim Retrospective/historical data to
make predictions with regard to
budget, future enrollments &
resource allocation on
institutional level
Plus real-time data for real-time
interventions
Default Forgetting Remembering
Personal
identifiers
Anonymised, aggregated data Plus re-identifiable data
Personal/ised data
Oversight/
data
governance
Broad institutional oversight.
Ethical Review Board (ERB)
approval for research purposes
Approval, oversight and
governance highly complex and
contested
39. (1)
Humans
perform the
task
(2)
Task is
shared with
algorithms
(3)
Algorithms
perform task:
human
supervision
(4)
Algorithms
perform task:
no human
input
Seeing Yes or No? Yes or No? Yes or No? Yes or No?
Processing Yes or No? Yes or No? Yes or No? Yes or No?
Acting Yes or No? Yes or No? Yes or No? Yes or No?
Learning Yes or No? Yes or No? Yes or No? Yes or No?
Danaher, J. (2015). How might algorithms rule our lives? Mapping the logical space of algocracy. [Web log post]. Retrieved from
http://philosophicaldisquisitions.blogspot.com/2015/06/how-might-algorithms-rule-our-lives.html
Human-algorithm interaction in the collection, analysis and
use of student data: What are the (ethical) issues?
2
42. Griffiths, D. (2017, September). An Ethical Waiver for Learning Analytics?. In European Conference on Technology
Enhanced Learning (pp. 557-560). Springer, Cham.
4
43. Willis, J. E., Slade, S., & Prinsloo, P. (2016). Ethical oversight of student data in learning analytics: A typology derived from a cross-
continental, cross-institutional perspective. Educational Technology Research and Development, 64, 881-901. DOI: 10.1007/s11423-
016-9463-4 http://link.springer.com/article/10.1007/s11423-016-9463-4
Who will provide oversight over the ethical
issues in learning analytics?
An interpretative multiple-case study: Indiana University, Open University (UK) and
the University of South Africa (Unisa)
44. “Ethics are the mirror in which we
evaluate ourselves and hold ourselves
accountable” (emphasis added).
Holding actors and humans accountable
still works “better than every single
other system ever tried” (Brin, 2016)
Ethics and accountability
48. Source credit: https://www.open.ac.uk/students/charter/sites/www.open.ac.uk.students.charter/files/files/ecms/web-content/ethical-
use-of-student-data-policy.pdf
Principle 1: Learning analytics is an ethical practice that should align with core
organisational principles, such as open entry to undergraduate level study.
Principle 2: The OU has a responsibility to all stakeholders to use and extract
meaning from student data for the benefit of students where feasible.
Principle 3: Students should not be wholly defined by their visible data or our
interpretation of that data.
Principle 4: The purpose and the boundaries regarding the use of learning
analytics should be well defined and visible.
49. Source credit: https://www.open.ac.uk/students/charter/sites/www.open.ac.uk.students.charter/files/files/ecms/web-content/ethical-
use-of-student-data-policy.pdf
Principle 5: The University is transparent regarding data collection, and will
provide students with the opportunity to update their own data and consent
agreements at regular intervals.
Principle 6: Students should be engaged as active agents in the implementation of
learning analytics (e.g. informed consent, personalised learning paths,
interventions).
Principle 7: Modelling and interventions based on analysis of data should be sound
and free from bias.
Principle 8: Adoption of learning analytics within the OU requires broad acceptance
of the values and benefits (organisational culture) and the development of
appropriate skills across the organisation.
50. A long time ago in a far-off galaxy made of
academic conferences…
Source credit: https://www.jisc.ac.uk/guides/code-of-practice-for-learning-analytics
Guiding principles
1. Responsibility
2. Transparency and consent
3. Privacy
4. Validity
5. Access
6. Enabling positive interventions
7. Minimising adverse impacts
8. Stewardship of data
52. Processes
Inter & intra-
personal
domains
Modalities:
• Attribution
• Locus of control
• Self-efficacy
Processes
Modalities:
• Attribution
• Locus of
control
• Self-efficacy
Domains
Academic
Operational
Social
TRANSFORMED INSTITUTIONAL IDENTITY & ATTRIBUTES
THE STUDENT AS AGENT
IDENTITY, ATTRIBUTES, HABITUS
Success
THE INSTITUTION AS AGENT
IDENTITY, ATTRIBUTES, HABITUS
SHAPING CONDITIONS: (predictable as well as uncertain)
SHAPING CONDITIONS: (predictable as well as uncertain)
Choice,
Admission
Learning
activities
Course
success
Gradua-
tion
THE STUDENT WALK
Multiple, mutually constitutive
interactions between student,
institution & networks
F
I
T
FIT
F
I
T
FIT
Employ-
ment/
citizenship
TRANSFORMED STUDENT IDENTITY & ATTRIBUTES
F
I
T
F
I
T
F
I
T
F
I
T
F
I
T
F
I
T
F
I
T
F
I
T
Retention/Progression/Positive experience
Subotzky, G., & Prinsloo, P. (2011). Turning the tide: a socio-critical model and framework for improving student success in open distance
learning at the University of South Africa. Distance Education, 32(2): 177-19.
55. Imagecredit:https://pixabay.com/en/binary-code-man-display-dummy-face-1327512/
Student data sovereignty
Student data are not something separate from
students’ identities, their histories, their beings.
Data are an integral, albeit informational part of
students being.
In the light of the view that data are not
something students own but rather who they
are; what are we assuming when we say we
‘collect’ their data?
E.g. Floridi, L. (2005). The ontological interpretation of informational privacy. Ethics
and Information Technology, 7(4), 185-200.
6
59. Institutions have not only a fiduciary duty to collect, analyse
and use student data, but also a moral duty.
We never ‘own’ the data – students don’t ‘throw off’ data as
they engage with their materials, their peers, their
instructors. Their data are so much more than ‘digital
footprints’. We need to recognise their data sovereignty.
An ethical approach to collecting, analysing and using
student data should be a powerful counter-narrative to
some of the current the dominant discourses driven by
economic interests.
60. THANK YOU
Paul Prinsloo (Prof)
Research Professor in Open Distance Learning
(ODL), College of Economic and Management
Sciences, Samuel Pauw Building, Office 5-21,
P.O. Box 392, Unisa, 0003, Republic of South
Africa
T: +27 (0) 12 433 4719 (office)
prinsp@unisa.ac.za
Skype: paul.prinsloo59
Personal blog:
http://opendistanceteachingandlearning.wordpress.com
Twitter profile: @14prinsp
Image credit: https://pixabay.com/en/figure-dragon-wing-face-ceramic-3124002/