2. The OU Leading the Way!
• World leaders in applying learning analytics to benefit
disabled students in particular
– 19,000+ disabled students a key driver in this
• Main Actors:
– Martyn Cooper (IET); Rebecca Ferguson (IET); Annika Wolff (KMi)
• Forums:
– https://www.linkedin.com/groups/Learning-Analytics-Accessibility-
4369422?trk=my_groups-b-grp-v
– http://martyncooper.wordpress.com/
3. LA for
Accessibility
• Hypothesis – if there is a
module or learning activity
on which disabled students
consistently doing worse
than non-disabled students
– it might indicate
accessibility deficits
• Challenges:
• Data issues
• Disabled students often do
better than non-disabled
• Variable results comparing
successive presentations
• Subject differences
• Even if it indicates where a
deficit is it does not say
anything about what it is -60.0
-40.0
-20.0
0.0
20.0
40.0
60.0
80.0
100.0
120.0
1
62
123
184
245
306
367
428
489
550
611
672
733
794
855
916
977
1038
1099
1160
1221
1282
1343
1404
Nondisabled - Disabled % Complete
Disabled-Nondisabled %
Complete
Number +ve 923 69%
Number -ve 415 31%
5. “Clickometrics”
• The VLE records every
“click” of every student
– So know:
• How long they have been
online
• How long engage with a given
activity
• Which activities, forums etc.
they interacted with
• Huge volume and velocity
of data – but what can we
measure with it?
• KMi has been doing work on
retention using this data
– Level of “clicking” not a measure
of student in trouble but changes
in level a good predictor
– Use as a launch for intervention
[Wolff, A., Zdrahal, Z., Herrmannová, D. and
Knoth, P. (2013) Predicting Student Performance
from Combined Data Sources, in eds. Alejandro
Peña-Ayala, Educational Data Mining:
Applications and Trends, 524, Springer]
6. Data Issues
• Noise
–Disabled students subject to all the other factors that
might impact on completion or pass rates
• Disabled students are far from a homogenous group
• User modelling is currently very crude
• Accessibility would be better assessed if we knew the
nature of the content but this not readily available without
going back to Module Teams
• Data not always robust and technical issues getting to it live
7. User Modelling
• Learning analytics for
disabled students is only
viable because the
university knows which
students declare a
disability
• <disability flag? (binary)
• <disability type>
(12 medical model categories)
• Functional models
– Could be based on IMS
AccessForAll 3.0
– Model access needs not
personal traits
– Needs students to complete
web form
8. <Disability type>
• Based on data collected
for HESA
• Does not map to
accessibility needs and
preferences
No. Descriptor
1 Sight
2 Hearing
3 Mobility
4 Manual skills
5 Speech
6 Specific learning difficulty e.g.
dyslexia
7 Mental health
8 Personal care
9 Fatigue/pain
10 Other
11 Unseen disability e.g. diabetes,
epilepsy, asthma
12 Autistic Spectrum Disorder
12. Accessibility of Learning Analytics
• If learning analytics (in
general and special
support for disabled
students) is going to be
of use then dashboards
must be accessible
– General accessibility
guidelines apply (WCAG
2.0 / OU Guidelines)
13. Institute of Educational Technology
The Open University
Walton Hall
Milton Keynes
MK7 6AA
www.open.ac.uk