The document discusses visions for the future of learning analytics based on a presentation given by Rebecca Ferguson. It outlines several potential futures for learning analytics, including learners being monitored by their learning environments, learners' personal data being tracked, and learners controlling their own data. It also discusses various challenges regarding ethics, regulation, validity, and affect that will need to be addressed for learning analytics to achieve its potential while avoiding negative consequences. The overall message is that learning analytics show promise to improve education if developed and applied carefully and ethically with student well-being and consent as top priorities.
1. Rebecca Ferguson, The Open University, UK
LASI Asia, 19 September 2016
Visions of the future:
what is on the horizon
for learning analytics?
2. The Open University (OU)
• The Open University: largest in UK
• Distance university
• Making use of big data for 45 years
• Informal learning: iTunes, YouTube…
• MOOCs on FutureLearn,
OpenLearn and elsewhere
• Learning analytics research and events
• Learning Analytics Community
Exchange
• LAEP – for European educational policy
2
www.laceproject.eu
laepanalytics.wordpress.com
4. In order to make
sensible decisions,
we need to look
ahead and plan
for the future
5. Priority areas for education and training
5
Open and innovative education and training, fully embracing
the digital era.
Strong support for teachers, trainers, school leaders and
other educational staff
Relevant and high-quality knowledge, skills and
competences developed throughout lifelong learning
Focus on learning outcomes for
employability, innovation, active citizenship and well-being
and inclusive education, equality, equity, non-discrimination
and the promotion of civic competences.
6. Learning analytics help us
to identify and make sense
of patterns in the data
to improve our teaching,
our learning and
our learning environments
7. Priority areas for education and training
7
• Bringing together different sectors: higher education, schools & workplace learning
• Building networks that will outlive the project’s funding period
• Helping to develop learning analytics capability
• Creating and sharing resources
• Developing visions of the future and agreeing how to work towards them
11. Provocation 1:
Learners are
monitored by their
learning
environments
Just as in bioethics,
there are
fundamental
human factors at
play here.
Too much Big
Brother vision to
be appealing.
I think it is a promising
line of work for
learning analytics. I
think there will be
many advances in the
use of sensors to
identify aspects that
can be applied to this
vision.
13. Provocation 2:
Learners’ personal
data is tracked
Wearable sensors are
already present, but in
the next future they
must be improved,
especially for health
purposes, such as
diabetes monitoring or
cardiovascular diseases
prevention.
Quantified self strongly
builds on reflection and
self-organisation. This
is good and feasible.
A reliable evidence base
for the effectiveness of
these measures and some
sort of safety control to
prevent irresponsible
recommendations.
15. Provocation 3:
Analytics are
rarely used
There should be strong
discussion about the
Ethical concerns that
apply to each approach
to Learning Analytics
the research community
must concretely show the
benefits of the use of data
to improve educational
outcomes.
Focus less on analytics to
automate, but rather to be
student-centered, to inform the
learner and improve their
personal practice. Once analytics
provides real, tangible value to
the learner, you begin to both
alleviate concerns about privacy
AND develop faith that if you can
give control of data back to the
learners, they will opt-in to
learning analytics.
17. Provocation 4:
Learners control
their own data
Absolutely
essential.
Organisations
must not rely on
the data, but try
to build trust
with their
learners so that
learners see the
benefit of
sharing.
Can we assume
that "data
owners" know
what to do with
their data? -
Some people
can't even
manage the
money in their
pockets.
Users should be entitled to
know how their data are
interrogated and used and
for what reasons. This
should be made explicit
and easy to understand
Greatly limits our ability to
effectively use learning
analytics to improve
learning for ALL students.
19. Provocation 5:
Open systems are
widely adopted
Define standards for the information
exchange - Define standards to
exploit the information stored -
Define guides about what
information is useful - Define and
publish different models to
represent the information
align with corporate
and stop thinking that
academic is leading!
A multi-pronged approach
involving IT services, national
organizations like JISC and
ministry of education and
tertiary institutions would all
need to be involved.
21. Provocation 6:
Learning analytics are essential tools
Very little credible research has
demonstrated any real large-scale
benefits to learners or institutions.
Learning is not only about
success is about learning from
failure. So, yes, I think that it
would be desirable that the
prediction rates are very high,
however, I don't think it is the
final goal.
What is needed is support
for reflection, discussion
and debate on the purpose
of it all, especially to curb
the excesses of those that
see learning as something
teachers do to students. We
need to nurture rich,
reflective communities
23. Provocation 7:
Analytics help learners
make the right choices
From bitter
experience, I'm aware
of the very slow pace
of institutional change
companies will sell
politicians on the
budget savings and
lead us here.
With learning there is an
element that learning is a
process – reducing it to
simply outcomes that can be
measured is dangerous
limited as ignores learning
through interactions with
both peers and the wider
social environment.
25. Provocation 8:
Analytics have largely
replaced teachers
autonomy begets engagement,
motivation, persistence, relevance.
The collective is as
important as the
individual: it is not
just about how *I*
learn but how *we*
learn.
Self-directed learning can let students
improve a lot according to their needs.
But they also need the instructors to
guide them when they are confusing
and frustrated during the learning
process.
26. Pedagogy We have a social duty to
facilitate and provide
opportunities for learners
to achieve their full
potential
• Why do we educate people?
• How do people learn?
• What pedagogic outcomes are we
trying to achieve?
• How can we measure those outcomes?
Learning is not only about
success – it is about
learning from failure
there is a time for
learners to be confronted
in order for transformation
and growth to occur
We need to nurture
rich, reflective
communities in both
teaching and learning
27. Smart houses, wearable
technology, the Internet of
Things and face
recognition are
increasingly part of
everyday life
Hard to believe that there
will be enough processing
power to do this, but I
guess people always say
that when something is
ten years away
A new government
authority that acts as a
trusted clearing house for
data and analytics
Complexity
• How can we understand the
internal process of learning by
measuring external actions?
• How do we engage a wide
range of stakeholders?
• How do we process huge
amounts of data from diverse
sources?
28. Ethics
• We need some form of
regulation in this area
• Control of data has
ethical implications
• Encourage awareness of
how data are used and
how analytics function
• Focusing on data as a
valuable commodity can
lead to unethical
practices
The key is to establish
the notion that each of us
own our own data: the
companies do not
institutional rules and
regulations must exist
and should meet
certain criteria
As long as the
data is
anonymous
data should be
allowed to be
used in these
kinds of
applications
without any
consent
29. One of the purposes of LA is to
empower the teachers to
provide better learning for the
individual learners
It is even worse to put
that control in the hands
of system designers and
programmers, thus
embedding their
assumptions and beliefs
• Who should control the data?
• Who should control the learning
and teaching process?
• Who sets goals for learners
and teachers?
• Who needs to understand
the analytic process?
Power
if tracking and monitoring
are used to foster and
support education and
learning, it might be
desirable. If it is used to
monitor and control and
to enforce power it is not
desirable
30. drawing on previous
legislation in the areas of
privacy, child protection,
data protection,
consumer protection, and
the use of personal data
in medical research
It must be handled as a human
right in the 21st century that
every single person should have
the power to decide, when + how
+ for what purpose + for which
timeframe + ... his/her personal
data can/cannot be used
• Need to regulate protection,
ownership and storage of data
• Need new policies on
education, ethics, privacy and
assessment
• Need to decide how this
regulation is developed and
enforced
Regulation
31. Very little credible
research has
demonstrated any real
large-scale benefits to
learners or institutions
The use of LA
applications in real
practice has be
conscious of the
limitations of any
analysis, and apply
them in a way that is
coherent with the
limitations of the
approach
we MUST be willing to
unpack the algorithms.
Academics are extremely
unlikely to accept 'black box'
predictive tools - it goes
against the very principles of
critical thought
Validity
How can we be sure that the
results generated by learning
analytics are valid, reliable
and generalisable?
32. Affect
• Bear in mind what
engages and motivates
teachers and learners
• Be aware that there is
discomfort and unease
about various aspects of
learning analytics
the real fuel of
Learning is
motivation and
volition, which you
cannot capture with
external sensors
I might be an alarmist, but there
is too much at stake: from
developing an underclass of
limited-dimension robiticized
learners, to propaganda-fed
righteous fanatics, an
automated, corrupted learning
environment puts us on a path
to an Orwellian future
autonomy begets
engagement, motivation,
persistence, relevance
34. 34
Slides online at www.slideshare.net/R3beccaF
Rebecca Ferguson @R3beccaF
http://r3beccaf.wordpress.com/
Notes de l'éditeur
Introducing the talk
The Open University (the picture is of the Jennie Lee building where IET is based)
The Learning Ananlytics Community Exchange (LACE) project, funded by Europe
Learning Analytics for European Education Policy (LAEP) project, funded by Europe
Both are focused, to some extent on the future of learning analytics
When we plan education and develop educational technology, we need to bear in mind that the world is changing
Education is not about training people for existing jobs – it is about equipping them with the skills to take onlife as a whole
We need to draw on experience, but we also need to look ahead to see what we are preparing people for
If we do not look to the future, we can make big mistakes
Part of envisioning the future is setting out a plan for what we will be doing in the next few years.
These are the European priority areas for education and training
Today I am talking about a specific aspect of planning for and envisioning the future – learning analytics
This is what I mean by learning analytics.
We take the traces that learners and teachers leave behind them and use those traces to make sense of what they are doing, what they intend to do and what they could do better
The study that I am going to describe was run by the LACE project
These are some of the things that the LACE project has been doing recently
One of the LACE project activities was working to develop a shared view of the future of learning analytics
To do this, we used a Policy Delphi.
This is a research method that brings together the views of experts about what is likely to happen in the future.The aim is not to achieve a consensus but to explore the range of future possibilities
In order to do this, the LACE team developed eight ‘provocations’ – visions of how the future could be
We shared these through a survey with international experts and with people interested in learning analytics
Each was asked to comment on two or more visions / provocations in terms of whether they were feasible / desirable and what would need to happen to make them into reality
I shall briefly outline the eight provocations.
Do any of them align with your own work, or do you feel you are working towards a very different vision of the future?
Provocation 1 relates to a world in which almost anything a learner uses can be used to collect data about their activities
These are some of the responses to that vision. People saw how it could be connected with sensor technology and the Internet of Things. They also raised the issue of Big Brother watching over learners and controlling what they do
Provocation 2 deals not with external data but with internal data. Information about where students are looking, how they are reacting to stimuli, what their heart rate is. The picture shows the Mindlfex game in which the blue ball is controlled by the user’s brainwaves – which suggests we are moving towards being able to detect thought patterns
Respondents linked this to the notion of the ‘quantified self’, and to activity in the field of medicine. They called for a reliable evidence base, which is an idea found in many of the responses to diffferent visions.
Provocation 3 is a negative one from the point of view of learning analytics. It suggests that there will be so many problems and controversial stories that learning analytics are no longer used in ten years time. The image refers to the multi-million dollar inBloom project, funded by the Gates Foundation, which had to close due to srong opposition from parents.
Issues here of ethics, and of WHY the analytics are being developed and applied
Provocation 4 is concerned with who owns the data. Should it be owned and controlled by individual learners?
Divergent opinions here. Some people think learners should control their data and that organisations should make this possible and desirable. Others think that this will make the data unusable and that it just adds needless extra responsibilities to the work of students
Provocation 5 is to do with getting learning analytics, and their related systems, to talk to each other and to understand each other. It also ties in with building a developer community.
IN order for this open approach to be possible, a lot of work needs to be done at local and national levels.
Note: Jisc (formerly the Joint Information Systems Committee) is a United Kingdom not-for-profit company whose role is to support post-16 and higher education, and research
Provocation 6 sees the role for learning analytics increasing, so that all learners are supported by a mound of data
However, this requires thought about what it means to learn, how learning takes place, and how learning analytics support that process
Provocation 7 sees a positive role for analytics, with control remaining in the hands of the learners.
Although this sounds a positive future, respondents could see potential problems.
The final provocation sees analytics replacing teachers
Any teacher that can be replaced by a computer deserves to be. This is a rewording by David Thornburg of the original Arthur C Clarke quote (“Teachers that can be replaced by a machine should be.”)
Again, this prompts consideration of how learning takes place and how it can best be supported
Analysis of these responses to our Policy Delphi helped us to identify seven major themes, with associated questions and issues.
The first of these is pedagogy – a theorised approach to learning and teaching
Second theme is complexity – lots of this will be difficult to do, but there are ways to develop this work, and precedents on which to build
Ethics was a theme that came up in relation to all the provocations
Power is a theme that has been less considered in relation to learning analytics – although sociologists are already querying the uses and implications of big data
Regulation ties in with both power and ethics. If learning analytics are to work well, we need checks and balances in place
Validity is an increasing concern as wel move away from small pilot projects to large-scale implementation
And personal responses are also important. If people aren’t happy with the analytics or if they don’t trust the analytics, then problems arise.
This was linked to a recurrent minor theme of alienation