Rebecca eynon learning & interaction in moo cs

Conceptualising learning and
interaction in MOOCs
Rebecca Eynon
SDP
9th July 2014

What is a MOOC?
• Massive Open Online Courses
– Not just open content, but a way for people to
interact and participate with resources and each other
within a set period of time
• Two key types:
– xMOOCs - learning goals are predefined by instructor
and structured by the environment (coursera, udacity,
edx, futureLearn)
– cMOOCs - based on connectivist principles, learner led
and open structure (etmooc, connectivistmoocs.org)

The promise
• “There is one big thing happening that leaves me incredibly
hopeful about the future, and that is the budding
revolution in global online higher education. Nothing has
more potential to lift more people out of poverty — by
providing them an affordable education to get a job or
improve in the job they have. Nothing has more potential
to unlock a billion more brains to solve the world’s biggest
problems. And nothing has more potential to enable us to
reimagine higher education than the massive open online
course, or MOOC, platforms that are being developed.”
(Thomas Friedman, New York Times, 2013)

A reality check
• “We have entered a new era in higher education, one
which is rapidly drawing the halls of academe into the
age of automation. Automation — the distribution of
digitized course material online (....) is assumed to
improve learning and increase wider access. In
practice, however, such automation is often coercive in
nature — being forced upon professors as well as
students — with commercial interests in mind (...) It is
not a progressive trend towards a new era at all, but a
regressive trend, towards the rather old era of mass
production, standardization and purely commercial
interests.”
(Noble, First Monday, 1998)

Beyond the hype
• MOOCs as just one part of the complex ecology of
higher education, which need to be explored together
with the existing and historical characteristics and
functions of the sector.
• As Robins and Webster (2002) suggest, researchers
need “to develop a more sociologically grounded
narrative of change in higher education – one that is
aware of continuities, as well as transformations” (pp.6-
7).

What’s new?
• The opportunity to learn openly available content together
– Communication via discussion forums
• The scale
– Teaching challenges (diversity of individuals)
– Methodological challenges (understanding behaviour and
experience in a “noisy” environment)
• A shift
– Increasing numbers of semi-formal “crowd-like” spaces where
learning can occur
– Methodological innovation – but area should be a specialism of
educationalists as well as data scientists

Overall goal
1. Develop profiles that reflect the different
ways and reasons that people interact with
one another in MOOCs
2. How these interaction profiles are related to
learner characteristics and course outcomes

Research approach
• Case study of one MOOC from Coursera with strong emphasis on
encouraging interactions between learners
• Use data from this MOOC to develop typology, then test on wider
data sets
• Mixed methodology
– Visualisation of posts and views, SNA, in-depth interviews, pre and
post surveys, qualitative observations & content analysis
– Moving between the quant and qual methods to keep refining the
model
• Other members of the team: Nabeel Gillani, Taha Yasseri, Isis Hjorth

Use of mixed method design
• Working within a pragmatic paradigm
• Three levels of data
– structural description (patterns of interactions);
– thin descriptions, which note the content of the
interaction
– thick description, to provide rich context and convey the
meaning of events to those who participated in them
(Welser et al., 2008)
• Linking methods at three different levels can be very
valuable particularly for learning

Learning is not just behaviour
• The process of learning unites a cognitive, an
emotional and a social dimension into one
whole. It combines a direct or mediated
interaction between the individual and its
material and social environment with an
internal psychological process of acquisition.
(Illeris, 2007: 227)
10

Quick overview
Theoretical schools
• Behaviourism
• Cognitivism
• Cognitive constructivism
• Social constructivism
• Socio-cultural
Technology
• Teaching machine
• Enhancing info processing
• Object to think / reflect with
• Supporting mediation
• As part of a wider social
system / community

The course
• 6 week course
• 4-6 hours per week
• Assignments
– Required: Weekly quizzes, final strategic analysis
assignment (evaluated via peer-assessment)
– Optional: Discuss business cases in the discussion forums
• Multiple sub-forums
– Final project, cases, lectures, readings, study groups,
questions for professor, technical feedback, course
material feedback

Participation
• 90,000 signed up
• 49,682 used the lecture videos at least once
• 20,082 submitted at least one quiz
• 4,445 posted at least once in discussion forums
(15,000 posts, 150,000 views)
• 2,208 received >= 70%
• What does it means to “participate” in these
settings?
13

Demographics of participants
• Nearly 40% of respondents were between 25 and 34
• Over 85% had at least a Bachelor’s degree, and approx
40% had a Master’s or professional degree
• Overall, this study supports the general finding that
well-educated people from the western world are most
likely to take MOOCs and pass the course
• Do MOOCs widen participation?

Patterns of interaction
• A framework for forum analysis
– Analyse interactions according to sub-forums –
justified by low participation overlaps between
sub-forums
• Forum activity is ‘bursty’, with most activity
occurring earlier in the course

Social network analysis of forums
)

Exploring the network
• Decisions about what data to include
– we cut by sub-forum, but what about time?
• What do we consider as a "tie" between two learners?
– two learners were connected if they co-participated in at
least one discussion thread (undirected)
• Which ties between learners are meaningful interactions and
not just “noise”?
– we generated possible communication networks based on
the trends in the observed network and tested for the
likelihood that the link was there by chance
)

Community structure in the network
• Which ‘critical set’ of learners is responsible
for potential information flow in a
communication network - what happens to
online discussions if the learners comprising
this set were removed?
• Video
19

So….
• Forums mostly harbour crowds, not
communities, of learners characterised by
weak ties
• Learning seems more likely to be occurring in
particular sub-forums (cases and final
projects)

Qualitative interviews
• How do people experience learning in
MOOCs?
• 30 interviews
• Key issues: sampling

MOOC learner types
• Key drivers
– Problem solving
– Professional profiling
– Lifelong learning
– Connecting and accreditation
• Approach to learning and role of forums differ
significantly between groups

MOOC learner type 1 [Problem
solving]
• Just in time learning
– MOOC engagement  Knowledge to assist specific decision-
making processes
– Key barriers: Time management/other work obligations
• Crowd of no significance for learning
– Learning as an individual pursuit
• Example: Carney, early 50s, Ireland, Masters’ degree
– We were trying to come up with a new idea for how we position
ourselves in the market, [...] and our group president came from
a strategy background, so I thought [...] we’d be best to talk in
his language and concepts he’s used to, so that at least he’d see,
that you know, we’re on the same track.

MOOC learner type 2 [Professional
profiling]
• Just in case learning
– MOOC engagement  Skill development for future
employment
– Key barriers: Time management/other work obligations
• Crowd of limited relevance for learning;
– May occasionally serve as information source
• Example: Julia, late 20s, England, masters’ degree
– The courses I’m taking more seriously, are the ones that
push me a bit more towards my career goals.
– I’m using it the MOOC quite selfishly, I suppose.

MOOC learner type 3 [Lifelong
learning]
• Life-long learning
– MOOC engagement  Extension of traditional educational
experiences
– Key barriers: appropriate collaborative tools, study group logistics
• Crowd of high importance for learning
– Serve as pool for knowledge co-construction and networking (on- and
offline)
• Example: Lucas, mid-20s, Master’s degree, Spain
– I use the forum to connect with other students and set up a group or
tools in other platforms. […] MOOCs that imply interaction with other
students, are much more interesting […] this interaction gives a boost
to your motivation.
– For me this is about constant learning, and I actually plan to do
courses throughout my life.

MOOC learner type 4 [Connecting &
formal accreditation]
• Access to education
– MOOC engagement  Global outreach and connectivity
– Key barriers: Internet access/speed; learning skills/culture
• Crowd crucial for learning
– Crowd integrated part of MOOC learning experience; source of
knowledge.
• Example: Emengo, early 40s, Bachelors’ degree, Nigeria
– I like the forums, you learn a lot. And it encourages you to learn
a lot, people teach you – without necessarily telling you what
the solution is.
– The world is getting smaller (…) if I interact with these people
across-border, then my knowledge also somehow has to be
across-border.”

Connecting the micro & the macro

Content analysis
• All posts in the cases and final projects forum
– 6500 posts
– Each post on 5 dimensions:
• Knowledge construction
• Communicative intent
• Emotion
• Topic
• Relevance
28

Quantitative interaction profiles
• Bayesian Non-negative Matrix Factorization
(BNMF) to extract communities of learners based
on the nature of their online forum posts
– Cases (knowledge construction and communicative
intent)
– Final projects (communicative intent and topic)
• Relationship between these profiles and
geography, education, outcome, average views
and posts

Distinct interaction profiles
Cases
• Committed crowd engagers
• Discussion initiators
• Strategists
• Individualists
Final projects
• Instrumental help seekers
• Careful assessors
• Community builders
• Focused achievers
• Project support seekers

Linking the micro to the macro
• 23 of our interviewees had been classified into
at least one of the interaction profiles
31

Committed crowd engagers (quant)
• Contributed the most responsive dialogue acts, second
highest number of informative and elicitative statements
• Similar higher-order knowledge construction to groups 2
&3
• Read and posted the most of all four groups
• Most likely to pass the course
• Likely to be from Western countries
• A third had at least a Master’s degree
• It is reasonable to suggest that this group found the forums
an important part of their learning and used it as they
sought to formally pass the course.

Committed crowd engagers (qual)
• The forums are very important to me to engage with the
material (Elizabeth, Canada)
• Sometimes people were missing significant points and
that’s when I would tend to post more (James, US)
• I am not looking to connect to other people in the sense of
getting to know them personally (…) it was quite interesting
to think about it and then just read other people's opinions
and compare them with your own views (Arthur, Germany)

Committed crowd engagers (qual)
• Passing the course has no value for me (James,
US)
• What concerns me in general, is that watching
lectures and doing simple quizzes doesn't really
teach you that much. Real learning requires
putting in some effort, be it additional reading or
exercises, and applying what you have learnt to
real situations (Arthur, Germany)

Strategists (quant)
• Similar levels of higher-order learning to crowd engagers
• A greater proportion of argumentative statements
• Rarely had posts that reflected no learning
• Second most likely to pass the exam
• Tended to be similarly educated to crowd engagers
• They viewed and contributed to the forums the second
most often, but this was still significantly less than crowd
engagers
• Combined, these characteristics suggest that learners were
more strategic in their approaches, using the cases forum
only as needed to achieve their learning goals.

Strategists (qual)
• I love the Superstar lecturers (...) [in the forums] I usually
look for very specific issues which are of interest (Sherwin,
Philippines)
• If I don’t get any training, it looks also like I’m not up to
date (...) the certificate is very important (Manuel, Spain)
• If I’m taking a course, I intend to get an A. This aspect is
very important, and it’s what’s actually driving me to finish
all the material. So in those courses [where forum
participation is mandatory], I’ve used the forum, but
otherwise, [it is] very limited (Oliver, Canada)

Strategists (qual)
• I really enjoy collaborating with people in my
job (...) but I found it was easier to just read
the materials on my own; I didn’t feel the need
to leverage the community in order to
complete the work (Oliver, Canada)

Overall....
• Not bad, but not perfect fit – can use this to
improve the model
• In general, quant typology fits well with different
theoretical ways of thinking about learning
• But people’s experiences are not so neat
• Real problem of missing data points and lack of
qual data from certain groups
• Issue of changes over time not captured by quant
aspect
38

Next steps
• Improve the model
• Test on far larger data set
• Development of a set of quantitative indicators to be used
in future research and practice
– Real-time, dashboard
• Recognition of diverse learner needs, motivations and
digital inequalities need to be considered to support any
kind of widening participation agenda

Further info
• MOOC Research Initiative funded by the Bill and Melinda
Gates Foundation
• Project site
– http://www.oii.ox.ac.uk/research/projects/?id=121
• Contact and papers
– rebecca.eynon@oii.ox.ac.uk

Rebecca eynon learning & interaction in moo cs

More Related Content

What's hot (20)

Viewers also liked (10)

Similar to Rebecca eynon learning & interaction in moo cs (20)

Rebecca eynon learning & interaction in moo cs