Prof. Dr. Hendrik Drachsler
@hdrachsler
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
3
• Hendrik Drachsler
Professor Educational Technologies
& Learning Analytics
• Research topics:
Recommender Systems
Learning Analytics
Mobile devices
• Application domains:
Schools
HEI
Medical education
WhoAmI
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
12/03/18 3Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
@HDrachsler, #
Slide 5 / 29 June
1. Definition of
trust and LA
2. Fears of
Learning
Analytics
Lecture structure
5. Technical
approaches
towards Trusted
Learning Analytics
3. Humboldt-ian
Educational
Model
4. New demands
from GDPR 2018
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
Trust is about a firm belief in the
reliability, truth, or ability of someone
or something.
A trustful relation is mutually based on
• openness
• truth
• reliability
• integrity
• belief
• faith
• freedom of suspicion
A definition of Trust
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
Picture by Terry Johnston
https://www.flickr.com/photos/powerbooktrance/466709245/
Greller, W. & Drachsler, H. (2012). Turning Learning into Numbers. Toward a Generic Framework for
Learning Analytics. Journal of Educational Technology & Society, 15(3), 42–57.
http://ifets.info/journals/15_3/4.pdf
A definition of Learning Analytics
@HDrachsler, #
Slide 8 / 29 June
1. Definitions
2. Fears of
Learning
Analytics
Lecture structure
5. Technical
approaches
towards Trusted
Learning Analytics
3. Humboldt-ian
Educational
Model
4. New demands
from GDPR 2018
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
Keynote Neil Selwyn @ LAK 2018, Sydney, Australia
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
Keynote Neil Selwyn @ LAK 2018, Sydney, Australia
Learning Analytics has a trust problem …
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
Keynote Neil Selwyn @ LAK 2018, Sydney, Australia
… because Learning Analytics
has the potential of becoming a
high stakes assessment.
Black box vs. White box
Open algorithms
Transparent indicators
No automated decisions
Full access to data
Knowing who accesses your data
Unknown algorithms
Unknown data collection
Automated decisions
No access to raw data
No control who uses it
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
Examples why people don’t trust
14
• $100 million investment
• Aim: Personalized learning in public schools, through data & technology
standards
• 9 US states participated, in 2013 data about millions of children
have been stored
Ignoring the fears and public
perception of the application of
analytics can lead to a lack of
acceptance, protests, and even failure
of entire LA implementations.
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
Examples why people don’t trust
@HDrachsler, #
Slide 15 / 29
1. Definitions
2. Fears of
Learning
Analytics
Lecture structure
5. Technical
approaches
towards Trusted
Learning Analytics
3. Humboldt-ian
Educational
Model
4. New demands
from GDPR 2018
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
Humboldt-ian Model
Designed an education system …
• based on unbiased knowledge
• combining research and teaching
• allowing students to choose their own
course of study
• develop autonomous reflected individuals
• education beyond vocational training only
• comprehensive general learning
• cultural knowledge
Picture cc-by: https://en.wikipedia.org/wiki/Alexander_von_Humboldt
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
Humboldt-ian Model in the age of Big Data
Designed an education system …
• based on unbiased knowledge
• combining research and teaching
• allowing students to choose their own course of
study (personalization)
• develop autonomous reflected individuals (self-
regulated learning)
• education beyond vocational training only
• comprehensive general learning
(lifelong learning)
• cultural knowledge
(formative feedback vs. summative assessment)
@HDrachsler, #
Slide 17 / 29
1. Definitions
2. Fears of
Learning
Analytics
Lecture structure
5. Technical
approaches
towards Trusted
Learning Analytics
3. Humbold-tian
Educational
Model
4. New demands
from GDPR 2018
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
“the biggest change to data
protection law for a
generation”
Elizabeth Denham,
Information Commissioner
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
GDPR 2018
GDPR 2018
• Right to be informed
• Right of access
• Right to rectification
• Right to erasure
• Right to restrict processing
• Right to data portability
• Right to object automated decision making
Do your Learning Technology systems
support these rights?
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
The GDPR may apply to an
institution even if it does not
have any physical presence in
the EU.
Interaction with EU citizens in a
way that results in data being
kept on them appears to trigger
the applicability of the law.
The law appears to apply to
study abroad programs as well.
GDPR 2018
21
Some things are already on its way
http://www.open.ac.uk/students/ch
arter/essential-documents/ethical-
use-student-data-learning-analytics-
policy#
https://www.jisc.ac.uk/sites/default/files/jd
0040_code_of_practice_for_learning_analy
tics_190515_v1.pdf
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
22
Drachsler, H. & Greller, W. (2016). Privacy
and Analytics – it’s a DELICATE issue. A
Checklist to establish trusted Learning
Analytics. 6th Learning Analytics and
Knowledge Conference 2016, April 25-29,
2016, Edinburgh, UK.
Online at:
http://www.laceproject.eu/ethics-privacy/
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
Some things are already on its way
Some things are already on its way
http://www.sheilaproject.eu
Yi-Shan Tsai, Pedro Manuel
Moreno-Marcos, Kairit
Tammets, Kaire Kollom, and
Dragan Gašević. 2018. SHEILA
policy framework: informing
institutional strategies and
policy processes of learning
analytics. In Proceedings of the
8th International Conference on
Learning Analytics and
Knowledge (LAK '18). ACM, New
York, NY, USA, 320-329. DOI:
https://doi.org/10.1145/317035
8.3170367
@HDrachsler, #
Slide 24 / 29
1. Definitions
2. Fears of
Learning
Analytics
Lecture structure
5. Technical
approaches
towards Trusted
Learning Analytics
3. Humbold-tian
Educational
Model
4. New demands
from GDPR 2018
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
Trusted Learning Analytics Infrastructure
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
• TRU-LA first GDPR 2018
conform Big Data
Infrastructure
• Joined project with
GU und OU
• Among ‘traditional‘ learning
data we also aim to collect
multimodal data.
Trusted Learning Analytics Infrastructure
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
Trusted Learning Analytics in action
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
• Context
– Online course on sustainable development with 60 students from 11
EU universities
• Key components
– students (and staff) of different nationalities and backgrounds
(cultural and disciplinary)
– collaboration between geographically distributed students
• RQ:
– How can we best integrate learning analytics and group awareness
for students in CSCL learning environments to optimize group
dynamics and collaborative learning?
Trusted Learning Analytics in action
Cumulative View Periodic View
Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
Trusted Learning Analytics in action
Reciprocal Privacy Relationship
Scheffel, M., Drachsler, H., Kreijns, K., de Kraker, J., & Specht, M. (2017). Widget, widget as
you lead, I am performing well indeed! - Using results from a formative offline study to
inform an empirical online study about a learning analytics widget in a collaborative learning
environment LAK’17. Vancouver, Canada: ACM.
The Widget
The Widget
The Widget
The Widget
And now guess, how many of the 60 students
wanted their data hidden?
@hdrachsler drachsler@dipf.de
35
Many thanks for your attention!
Questions now or later:

Trusted Learning Analytics

  • 1.
    Prof. Dr. HendrikDrachsler @hdrachsler Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 2.
    3 • Hendrik Drachsler ProfessorEducational Technologies & Learning Analytics • Research topics: Recommender Systems Learning Analytics Mobile devices • Application domains: Schools HEI Medical education WhoAmI Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 3.
    12/03/18 3Adelaide |12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 5.
    @HDrachsler, # Slide 5/ 29 June 1. Definition of trust and LA 2. Fears of Learning Analytics Lecture structure 5. Technical approaches towards Trusted Learning Analytics 3. Humboldt-ian Educational Model 4. New demands from GDPR 2018 Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 6.
    Trust is abouta firm belief in the reliability, truth, or ability of someone or something. A trustful relation is mutually based on • openness • truth • reliability • integrity • belief • faith • freedom of suspicion A definition of Trust Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit Picture by Terry Johnston https://www.flickr.com/photos/powerbooktrance/466709245/
  • 7.
    Greller, W. &Drachsler, H. (2012). Turning Learning into Numbers. Toward a Generic Framework for Learning Analytics. Journal of Educational Technology & Society, 15(3), 42–57. http://ifets.info/journals/15_3/4.pdf A definition of Learning Analytics
  • 8.
    @HDrachsler, # Slide 8/ 29 June 1. Definitions 2. Fears of Learning Analytics Lecture structure 5. Technical approaches towards Trusted Learning Analytics 3. Humboldt-ian Educational Model 4. New demands from GDPR 2018 Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 9.
    Keynote Neil Selwyn@ LAK 2018, Sydney, Australia Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 10.
    Keynote Neil Selwyn@ LAK 2018, Sydney, Australia Learning Analytics has a trust problem … Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 11.
    Keynote Neil Selwyn@ LAK 2018, Sydney, Australia … because Learning Analytics has the potential of becoming a high stakes assessment.
  • 12.
    Black box vs.White box Open algorithms Transparent indicators No automated decisions Full access to data Knowing who accesses your data Unknown algorithms Unknown data collection Automated decisions No access to raw data No control who uses it Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 13.
    Examples why peopledon’t trust
  • 14.
    14 • $100 millioninvestment • Aim: Personalized learning in public schools, through data & technology standards • 9 US states participated, in 2013 data about millions of children have been stored Ignoring the fears and public perception of the application of analytics can lead to a lack of acceptance, protests, and even failure of entire LA implementations. Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit Examples why people don’t trust
  • 15.
    @HDrachsler, # Slide 15/ 29 1. Definitions 2. Fears of Learning Analytics Lecture structure 5. Technical approaches towards Trusted Learning Analytics 3. Humboldt-ian Educational Model 4. New demands from GDPR 2018 Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 16.
    Humboldt-ian Model Designed aneducation system … • based on unbiased knowledge • combining research and teaching • allowing students to choose their own course of study • develop autonomous reflected individuals • education beyond vocational training only • comprehensive general learning • cultural knowledge Picture cc-by: https://en.wikipedia.org/wiki/Alexander_von_Humboldt Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit Humboldt-ian Model in the age of Big Data Designed an education system … • based on unbiased knowledge • combining research and teaching • allowing students to choose their own course of study (personalization) • develop autonomous reflected individuals (self- regulated learning) • education beyond vocational training only • comprehensive general learning (lifelong learning) • cultural knowledge (formative feedback vs. summative assessment)
  • 17.
    @HDrachsler, # Slide 17/ 29 1. Definitions 2. Fears of Learning Analytics Lecture structure 5. Technical approaches towards Trusted Learning Analytics 3. Humbold-tian Educational Model 4. New demands from GDPR 2018 Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 18.
    “the biggest changeto data protection law for a generation” Elizabeth Denham, Information Commissioner Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit GDPR 2018
  • 19.
    GDPR 2018 • Rightto be informed • Right of access • Right to rectification • Right to erasure • Right to restrict processing • Right to data portability • Right to object automated decision making Do your Learning Technology systems support these rights? Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 20.
    The GDPR mayapply to an institution even if it does not have any physical presence in the EU. Interaction with EU citizens in a way that results in data being kept on them appears to trigger the applicability of the law. The law appears to apply to study abroad programs as well. GDPR 2018
  • 21.
    21 Some things arealready on its way http://www.open.ac.uk/students/ch arter/essential-documents/ethical- use-student-data-learning-analytics- policy# https://www.jisc.ac.uk/sites/default/files/jd 0040_code_of_practice_for_learning_analy tics_190515_v1.pdf Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 22.
    22 Drachsler, H. &Greller, W. (2016). Privacy and Analytics – it’s a DELICATE issue. A Checklist to establish trusted Learning Analytics. 6th Learning Analytics and Knowledge Conference 2016, April 25-29, 2016, Edinburgh, UK. Online at: http://www.laceproject.eu/ethics-privacy/ Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit Some things are already on its way
  • 23.
    Some things arealready on its way http://www.sheilaproject.eu Yi-Shan Tsai, Pedro Manuel Moreno-Marcos, Kairit Tammets, Kaire Kollom, and Dragan Gašević. 2018. SHEILA policy framework: informing institutional strategies and policy processes of learning analytics. In Proceedings of the 8th International Conference on Learning Analytics and Knowledge (LAK '18). ACM, New York, NY, USA, 320-329. DOI: https://doi.org/10.1145/317035 8.3170367
  • 24.
    @HDrachsler, # Slide 24/ 29 1. Definitions 2. Fears of Learning Analytics Lecture structure 5. Technical approaches towards Trusted Learning Analytics 3. Humbold-tian Educational Model 4. New demands from GDPR 2018 Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 25.
    Trusted Learning AnalyticsInfrastructure Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 26.
    • TRU-LA firstGDPR 2018 conform Big Data Infrastructure • Joined project with GU und OU • Among ‘traditional‘ learning data we also aim to collect multimodal data. Trusted Learning Analytics Infrastructure Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 27.
    Trusted Learning Analyticsin action Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit
  • 28.
    • Context – Onlinecourse on sustainable development with 60 students from 11 EU universities • Key components – students (and staff) of different nationalities and backgrounds (cultural and disciplinary) – collaboration between geographically distributed students • RQ: – How can we best integrate learning analytics and group awareness for students in CSCL learning environments to optimize group dynamics and collaborative learning? Trusted Learning Analytics in action
  • 29.
    Cumulative View PeriodicView Adelaide | 12.03.2018 | @hdrachsler | Digitale Learning Summit Trusted Learning Analytics in action
  • 30.
    Reciprocal Privacy Relationship Scheffel,M., Drachsler, H., Kreijns, K., de Kraker, J., & Specht, M. (2017). Widget, widget as you lead, I am performing well indeed! - Using results from a formative offline study to inform an empirical online study about a learning analytics widget in a collaborative learning environment LAK’17. Vancouver, Canada: ACM.
  • 31.
  • 32.
  • 33.
  • 34.
    The Widget And nowguess, how many of the 60 students wanted their data hidden?
  • 35.
    @hdrachsler drachsler@dipf.de 35 Many thanksfor your attention! Questions now or later:

Editor's Notes

  • #2 Who of you are university staff, and are there any students in the room? I gone share some very urgent issues with you guys. Actually, I more into more fancy stuff like multimodal learning analytics where we apply sensors and LA tech. to support learners to give presentations, how to dance, or provide feedback to meetings on various indicators. But today I need to talk with you guys about something very important that I call ‘Trusted Learning Analytics’ that is key for the rollout of LA in Europe.
  • #3 But first a few slides about myself
  • #4 Welten Institute Research Centre for Learning, Teaching and Technology Open University, Heerlen, The Netherlands Distance Education & Educational Sciences
  • #5 School of Computer Science Goethe University Frankfurt German Institute for international Educational Research (DIPF) Frankfurt am Main, Germany F2f teaching & computer science
  • #6 We go from very high level theoretical parts to very concrete early prototypes of trusted learning analytics. I really like to make tangible applications to experience how an thepretical approach materializes itself in a concrete mean and what are the pros- and cons of that mean in practice.
  • #8 Holistic approach towards Learning Analytics that is framing the Research Programme I’m chairing now.
  • #10 Neil Selwyn described an dystopia of learning analytics and mapped out the promises and problems of LA.
  • #14 A few years ago a research article published about the faebook emotions study that resulted in huge protests. Facebook conducted a research experiment that damaged the social networks relationship with users. Not all research has been compliant with the established ethical consensus.
  • #17 Who of you knows about Humboldt? Prussian philosopher (1769 – 1859) 1810 reformed the education system to humanist principles He got the tasks to setup a totally new education system in 1810 for the Prussian state at that time. Humboldt educational model goes beyond vocational training. In a letter to the Prussian king, he wrote: "There are undeniably certain kinds of knowledge that must be of a general nature and, more importantly, a certain cultivation of the mind and character that nobody can afford to be without. People obviously cannot be good craftworkers, merchants, soldiers or businessmen unless, regardless of their occupation, they are good, upstanding and – according to their condition – well-informed human beings and citizens. If this basis is laid through schooling, vocational skills are easily acquired later on, and a person is always free to move from one occupation to another, as so often happens in life.
  • #19 GDPR is partly framing the Humboldtian values into concrete rights of data subjects for the 21st / Big Data century
  • #23 Engelfriet, A., Jeunink, E., Manderveld, J. (2015).