This is the work presented at FutureLearn Academic Network Conference on 15.06.2015, in UK Open University. The presentation gives a brief explanation of analysis I have done to understand the importance of social network analysis in MOOCs through personalised MOOCs.
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
FLAN conference: MOOCs, personalisation, social networks
1. The State of the Art of
Personalised MOOCs
&
The Importance of Social
Learning Networks Analysis
Ayşe Saliha SUNAR
15.06.2015
2. My education journey
15 June 2015Ayşe Saliha Sunar @aysesCS
1
Turkey
BSc:
Mathematics
Japan
MSc:
Intelligent
Tutoring
Systems
United Kingdom
PhD:
Learning
Analytics &
Personalisation
& MOOCs
3. 15 June 2015Ayşe Saliha Sunar @aysesCS
2
MOOC Datasets management
Data Analysis
Curation:
Academic Literature (Mendeley)
Journalistic literature (Scoop.it)
Blog
Training
Publications
4. Motivation
15 June 2015Ayşe Saliha Sunar @aysesCS
3
Opportunity
Personalisation services by using learning analytics
Possible solution
Drawbacks
MOOCs produce large datasets from a large number of
learners.
Lack of customised education to meet each learner’s need
Low participation in online discussions & high attrition rate
5. The State of the Art of
Personalised MOOCs
15 June 2015Ayşe Saliha Sunar @aysesCS
4
6. CSEDU paper
7th International Conference on Computer Supported
Education, 23-25 May, 2015, Lisbon
15 June 2015Ayşe Saliha Sunar @aysesCS
5
Ayse Saliha SUNAR
Nor Aniza ABDULLAH
Hugh C DAVIS
Su WHITE
7. Focus of the paper
What is the nature of the literature on personalisation in
MOOCs?
What kind of issues are addressed by this literature?
What kind of techniques are applied for personalising
MOOC education?
Which directions could personalisation attempts in
MOOCs lead to?
15 June 2015Ayşe Saliha Sunar @aysesCS
6
8. Methodology: Critical &
Systematic Literature Review
15 June 2015Ayşe Saliha Sunar @aysesCS
7
Search for the keywords
“MOOCs personalisation”
and “adaptive MOOCs”
between 2011 and 2014
Google Scholar
The British Journal of
Educational Technology
American Journal of
Distance Education
Journal of Online Learning
and Teaching
ISI Web of Knowledge
IEEEXplorer
9. Growing attention to MOOCs
personalisation
15 June 2015
0
100
200
300
400
500
600
700
2011 2012 2013 2014
Total results for
"MOOCs
personalisation"
Total results for
"adaptive
MOOCs"
Relevant results
for "MOOCs
personalisation"
Relevant results
for "adaptive
MOOCs"
Ayşe Saliha Sunar @aysesCS
8
The total number of papers and relevant papers by the searches
10. Findings
Several techniques used for addressing several
issues
15 June 2015
Adaptive content presentation Poor pedagogical design
Adaptive navigation • Poor content engagement
• Lack of study skills
Adaptive feedback Poor feedback
Recommender system • Heterogeneity of learners
• High dropouts
• Excessive information on
discussion forums
Ayşe Saliha Sunar @aysesCS
9
11. Use of social network analysis
Researchers analysed enormous data of learners on
discussion forums and other social media tools to
answer:
Which social media tools are used often?
Who contributed more?
Who interacted with whom?
Is there any pattern of relations?
15 June 2015Ayşe Saliha Sunar @aysesCS
10
12. The Importance of Social
Learning Networks Analysis
15 June 2015Ayşe Saliha Sunar @aysesCS
11
13. Social networks on FL
Analysis of Develop Your Research Project MOOC (15
September – 5 November 2014 )
Dataset: Learners’ comment on the discussion boards
(15 September – 22 November 2014)
Focus of study:
How much did the learners contribute to online
discussions?
Did they sustain their contribution to online discussions?
Did recurrent interactions occur over the weeks?
15 June 2015Ayşe Saliha Sunar @aysesCS
12
14. Analysis of DYRP MOOC
This study has been
submitted as a full
paper to ICKM2015.
15 June 2015Ayşe Saliha Sunar @aysesCS
13
Week1
Week4
Week7
Week2
Week5
Week8
Week3
Week6
Learners’ interactions and the strength of their interactions in each week
15. Recurrent interactions in
DYRP
15 June 2015Ayşe Saliha Sunar @aysesCS
14
Only 19 peers
interacted multiple
times in different
weeks!
Those learners
joined all weeks
and completed the
course.
The recurrent interactions
16. Use of ‘following’ tab on FL
15 June 2015Ayşe Saliha Sunar @aysesCS
15
Do FutureLearn learners use
FOLLOW function efficiently?
Do they use FOLLOWING tab
on discussion board for active
communication with their
following list?
Is the following list usually
empty?
17. My PhD research
The Follow function on FutureLearn is very convenient
for promoting continuous friendships. But it needs to be
improved.
A recommender system for friend recommendation
could enhance learners’ interactions.
15 June 2015Ayşe Saliha Sunar @aysesCS
16
Learners who have continuous friendships could sustain their
interactions and most likely complete the course.
Hypothesis
18. Mendeley
15 June 2015Ayşe Saliha Sunar @aysesCS
17
Collection of paper on personalisation in MOOCs
https://www.mendeley.com/groups/4715311/mooc-personalisation
19. 15 June 2015Ayşe Saliha Sunar @aysesCS
@aysesCS
ass1a12@soton.ac.uk
18
Editor's Notes
I like to talk about my educational background first. CLICK I received my bachelor degree in mathematics in Turkey, CLICK then I went to Japan for master education in technology enhanced learning, specifically ITS. CLICK Shortly after, I have started my phd in the University of Southampton. Now, I am a 2nd year phd student. My general phd interests are learning analytics, personalisation and MOOCs.
Student’s studies output
MOOCs have some drawbacks. For example, the curriculum of MOOCs doesn’t meet every needs of learners. Also, low participation in the course is commonly observed.
However, since MOOCs produce a large amount of data, this is an opportunity for research to overcome those shortcomings.
I like to talk about the fundamental study I have completed.
We have done a critical literature review on personalisation of MOOCs. This paper is presented in CSEDU conference last month.
The focus of the paper is to understand how researchers discussed the personalisation in MOOCs and what kind of personalisation techniques they have applied.
We have searched two keywords in between 2011 and 2014 on these digital academic database.
According to the analysis results, the total number of retrieved paper sharply increase in 2013. Yet the related number of papers are quite low. I should express that most of the related paper were on Google Scholar and the related papers we consider here are not only peer-reviewed paper, but also from the grey literature for example working papers, project reports.
In the conducted and partly or fully implemented researches, several techniques used for addressing several issues such as adaptive content presentation, adaptive navigation and so on. Among the papers, the most commonly applied technique is recommender system and mostly used for fostering learners engagement in online discussion forums and helping them finding useful information and reducing high dropouts.
In those researches, researchers especially looking for the answer those question to propose a solution.
Next, I like to talk about my search and analysis on social learning networks in MOOCs.
I have analysed learners’ interactions in the Develop Your Research Project MOOC delivered by the University of Southampton. I have analysed learners’ comments and replies to someone else’s comment in order to answer these questions.
These are the learners’ interactions graphic. Interactions means that a learner actually post a reply to somebody’s comment. These graphics do not show the individual comments , but only replies. The z_axis shows the strength of interactions. I will not talk about how I calculated the strength of learner’s interactions. However, the strength is mainly based on the frequency of interactions. As you can see, the number of interactions is decreasing towards the end of the course. So do the strength of relationships.
This is the graphic showing the recurrent interactions. The numbers are anonymised learner’s IDs. Each mark identifies a recurrent interaction in a different week. If there are overlapped marks, that means those two learners got interacted again in another week. However, this kind of learners are only 19 peers. Also, we observed that those learners completed the course.
These results lead us to think about the functionality of FOLLOW functions. Do the FutureLearn MOOC learners follow each other? Do they check the following tab on the discussion boards? If they did, the recurrent interaction rate would have been higher.
Therefore, I have conducted my research on the hypothesis that Learners who have continuous friendships could sustain their interactions and most likely complete the course.
In order to enhance learners’ interactions, a recommender system for friend recommendation could be a very convenient method. This also could improve the Follow function on FutureLearn platform.
Lastly, this is the Mendeley group that you can find all literature on personalisation of MOOCs I have cited in my papers and more. You are very welcome to contribute to the group.
All feedbacks are appreciated. Thank you very much.