SlideShare utilise les cookies pour améliorer les fonctionnalités et les performances, et également pour vous montrer des publicités pertinentes. Si vous continuez à naviguer sur ce site, vous acceptez l’utilisation de cookies. Consultez nos Conditions d’utilisation et notre Politique de confidentialité.
SlideShare utilise les cookies pour améliorer les fonctionnalités et les performances, et également pour vous montrer des publicités pertinentes. Si vous continuez à naviguer sur ce site, vous acceptez l’utilisation de cookies. Consultez notre Politique de confidentialité et nos Conditions d’utilisation pour en savoir plus.
Going to be giving some background Talking about the need for social learning analytics Descrbing different types of analytic with examples from the OU
The UK ’ s largest University. Over 40 years old Runs its own courses and awards its own degrees 342 undergraduate modules, 60 CPD modules and 141 post-grad modules 43% have 1 A level or lower at entry Link this with ideas about MOOCs
59 million iTunes downloads Difference between formal and informal learning Shift towards online social learning
Developing SocialLearn - a social media space tuned for learning.
When it comes to analytics, the OU is well placed to work towards these aims of learning analytics We have a track record of research We have a lot of data to work with well over a million students in that 40 years, well over a million informal learners. Lots of researchers at our university are interested in how we can used this data to support learners and teachers. And we are interested in using analytics where the usual proxies for learning – completion and results – aren’t available or aren’t sufficient
Learning analytics can seem relatively new and small scale. For the last 20 years in England we have had a government driven analytics programme that results in the publication of school league tables in the national papers. Primarily academic analytics – designed for comparison of institutions But they become learning analytics because learners and teachers know about them and work with them We can take some positive messages from this
But also negative messages, and these you are familiar with in Australia. This is from the Morning Herald this week, on the NAPLAN tests I think part of the problem here is that these tests weren ’ t designed to solve a problem for teachers and learners, they were designed with government concerns in mind And they are not strongly driven by the learning sciences – in fact, they promote an instructivist approach to learning We always need to think about why we are developing analytics and who they will benefit
So, in that context, we want to support peopleto learn together. Not just in courses and cohorts, we are also interested in networks, communities and affinity groups. Two perspectives: 1. how can the individual learn more effectively in social contexts? 2. how can these groups function more effectively to support learning?
We need to adapt to an educational landscape where content expertise is no longer key; where we have to develop students who know how to learn, and are able to keep on learning, finding their own support and resources, even when things get tough
Some of the reasons we need social learning analytics: Social media are now ubiquitous, and support a lot of learning interactions Most of this is off the radar for formal education, there’s just too much of it, and learners often want to keep their social interaction separate from their formal education. Social learning analytics could provide tools to support learning in these situations Why this pic? – Google Plus, Twitter and Facebook being used for learning
Learners are faced with a deluge of information Just one Twitter hashtag can provide ideas, information, resources and support Social learning analytics could help to filter and recommend resources Could help to develop effective learning networks Why this pic? – the #phdchat community helping with filtering and recommendations
Knowledge, rather than land, labour or capital is now the key wealth-generating resource Constant change in society is now the norm We don’t know which knowledge and skills will be useful in the future, so we need ‘knowledge-age skills’ These included group-focused skills such as collaboration, communication and social responsibility Why this pic? - Two of the many knowledge-age skills frameworks
Sociocultural understanding of learning makes it clear that knowledge isn’t simply transferred to us There is an active process of knowledge construction The quality of the interaction round resources is important when making sense of the learning resources Need to be able to support students in the process of knowledge construction Why this pic? – Dragan in a learning analytics MOOC, the learning comes from the chat panel on the left, as well as from the presenter and his slides
Sociocultural understanding of learning makes it clear that learning is a social process It takes place using tools SLA could support learners to engage with these group tools, and in these different group environments Why this pic? – screenshot from ALT-C, using Internet, livestream, twitter stream while at F2F conference
MOOCs align with the OU values They are a growing global phenomenon – that relies to a greater or lesser extent, on the support provided by online social learning
We are drawing on a set of established research and tools. These are the ones we are currently exploring Network analytics – interpersonal relationships Discourse analytics – primary tool for knowledge construction Disposition analytics – these include learning relationships, and are developed through mentored conversations Content analytics – Resources have a social dimension – they are created, tagged, rated, reviewed, bookmarked Context analytics – giving access to the people, groups, social settings
Social network analytics are relatively well developed Several tools being developed to support them, including SNAPP and Network Awareness Tool Already in use, supporting both learners and teachers. We are interested not only in feeding back to individuals, but also to groups – what is working well, what do successful groups look like? For each set of learning analytics – there is a slide on description or possible uses, and one on how we are implementing it.
Here’s a form of network analysis we have been working on at the Open University, along with OUNL Taking an offline tool – NAT – and putting it online
The overall map is confusing, but you can filter it down by topic or by the type of tie
View the network for specific tags
Viewing by individual learner The types and amount of relationships
We currently have a lot of work at The Open University on discourse analytics From a sociocultural perspective, language and dialogue are crucial tools in the development of knowledge The work of Neil Mercer and his colleagues has shown that effective dialogue can be taught, and can significantly improve results
Here’s an example of data from an OU conference in Elluminate / Blackboard Collaborate We are interested in the synchronous text chat We started off with a two-day conference, and were looking for the times when people were engaging in exploratory dialogue in the chat
We looked for the characteristics of exploratory talk, and for the words and phrases that signalled these were taking place
Then we were able to run a simple search, which highlighted where these types of talk were taking place. For example, the sectionon the top left contains two of the indicators. This section also shows why other tools – that have been developed for more formal types of text, don’t work well on this type For example, we don’t always have full stops to mark the end of sentences, we have mistypes, and misspellings and abbreviations and emoticons But it’s also clear that this method doesn’t pick up all the examples – it is reliable, but it is incomplete So we started to work with computer linguistics
Framed it as a binary classification task – is a posting exploratory or is it non-exploratory? We developed a self-training approach that looked for unigrams, bigrams and trigrams – combinations of one, two or three words – that had a high probability of signalling exploratory or non-exploratory dialogue
Here we have a visualisation of a 2.5 hour session of the online conference For every ten postings, we have calcuated an average exploratory rate. These are shown with a timestamp Blue indicates non-exploratory dialogue, and red indicates exploratory dialogue. The longer the bar, the higher the probability So there are ten minutes at the beginning with a high amount of non-exploratory talk – and looking at these we see greetings There are also a lot of non-exploratory postings at the end, and these are mostly very similar Let’s look in more detail at the spike of exploratory talk just after 11am – highlighted in yellow here
At this finer granularity, the exploratory talk contains some sections that the computer clasfies as non-exploratory And you can see this needs some refinement – the classifier takes all URLs to be non-exploratory because sometimes they are off-topic Overall, though, there is a lot going on here, and this is a section it would be worth following up on
Where else might we employ this? A next move will be to use it in the OpenScience Lab to support students who are working together at a distance
Another example of discourse analysis at the OU In this case, the learners build semantic connections between contributions
This not only connects ideas, it gives an idea of the relationships between learners It also begins to identify key roles, such as ‘broker’
This, then, begins to move away from discourse analytics towards content analytics Once we know the semantic relationship between resources, we can make these available to learners
And a final example of discourse analytics FlashMeeting is a videoconferencing tool used on some of our course One of the views lets you see who has contributed when on the audio channel
This can be used for comparisons between sessions and between people
This takes into a personal domain – the characteristics of a good learner Validated over a decade – but also a list reflected in the Twitter stream Increase people’s capacity to learn, and their capacity to learn in a variety of situation
How dispositions are currently assessed
The dimensions of learning power, with a couple of descriptions
Ths is how learrners would see these dimensions reporesented
Understanding of these dimensions emerges through discussion
Different communities visualise the dimensions in different ways
We incorporated these dimensions within a Wordpress plug-in
The ELLI spider is already available as a form of learning analytic The social aspect appears in several ways. First, as the basis for a mentored conversation (see yesterday’s paper on mentoring analytics) Second, can we derive a useful profile from interactions on a social learning platform? This is what Shaofu is currently working on Why this pic? – ELLI pics from journal paper
This is a socialised analytic – recommendations are not necessarily social (note that this is not content analysis – which is a well-established method of analysis) What we are interested in here is whether we can make use of social information – ratings, recommendations, comments, bookmarks, to help learners to navigate information Recommendations aren’t necessarily based on interest – we could look for resources that challenge our views, that provide alternative perspectives
Social network reputation is not the same as domain expertise reputation
This is a model of the user-driven aspect of the Biodiversity Observatory and how it will help people with a casual interest into informal, and ultimately formal, learning about biodiversity.
Two approaches here. One takes your context as static – this might be your context when you are enrolled in a class, or working as part of a group The other takes a more dynamic approach and seeks to shift the possibilities available to you in your context – for example, you are walking down the street, and the analytics make you aware of a nearby resource or fellow learner (This view emerged from the MOBILEARN project)
Social Learning Analytics, Southern SoLAR Flare
Social Learning AnalyticsRebecca FergusonInstitute of Educational TechnologyThe Open University, UK Southern SoLAR Flare, Sydney
The Open University (UK)• Supported distance education• Quality open education at scale• Over 250,000 current students
Open learning 59,053,400 iTunes downloads 6,200,000 Science at the OU YouTube views OpenLearn 11,000 hours of viewing material 400,000 unique visitors per month Frozen Planet Have you studied with us?
Online social learning Why has someone sawn down half of the beautiful cedar tree outside my office window? I can’t find this out from a book, and I don’t know anyone with the precise knowledge that I am looking for. It is as I engage in conversations with different people that my understanding of what I see outside my window increases, and I learn more about the tree’s history, health, ecosystem and future possibilities.www.open.ac.uk/blogs/SocialLearnResearch/2010/01/13/what-is-social-learning-and-why-does-it-matter/
Learning analyticsDeveloping new tools for learners and teachersdrawing on experience from the learning sciencesintention of understanding and optimizingnot only learningbut also the environments in which it takes place
Implementing analytics • Aligned with clear aims • Huge and sustained effort • Agreed proxies for learning • Clear and standardised visualisation • Driving behaviour at every level Individual assessment within cohort
Negative perspective: NAPLAN There is massive pressure on schools and individual teachers to lift their school results. The logical consequence is to teach to the test … It is not about the students but it is all about the school … It does not reveal anything about the richness of your childs learning … Why is the ability to work in teams not included…? Sydney Morning Herald 27 November 2012
Social learning analyticsSocial learning analytics focus on how learnersbuild knowledge together in their cultural andsocial settings.In the context of online social learning, theseanalytics take into account both formal andinformal educational environments, includingnetworks and communities.
Why social learning analytics? (1)Changing environmentShift Happens:Karl Fisch
Why social learning analytics? (2)Social media Support learning-related reflection on interpersonal relationships and interactions
Why social learning analytics?Free and open content New Usefulideas resources Key hashtagsHelpful Support info networks Support the role of social networks in filtering and recommending resources
Why social learning analytics? (3)Living in the knowledge age Support learners to assess their progress in terms of knowledge-age skills
Why social learning analytics? (4)Sociocultural understandings Increase learner proficiency in the use of educational dialogue
Why social learning analytics? (5)Sociocultural understandings Enable learners to engage proficiently with a range of tools and social settings
Why social learning analytics? (6)MOOCs Open Translation MOOC http://bit.ly/OTcgjD Learning Design MOOC http://www.olds.ac.uk/
Social analytics: potential usesNetwork analyticsIdentify individuals who support my learningIdentify individuals with relevant interestsIdentify origins of conflictsIdentify groupings that could support learningProvide feedback to groups and group leaders
Network analyticsVisualising Social Learning in the SocialLearn Environment. Bieke Schreurs and Maarten de Laat (Open University, NL), ChrisTeplovs (Problemshift Inc. and University of Windsor), Rebecca Ferguson and Simon Buckingham Shum (The OpenUniversity, UK), SoLAR Storm webinar, The Open University, UK. http://bit.ly/UaFhbL
Social analytics: potential usesDiscourse analyticsThe ways in which learners engage in dialogueindicate how they engage with the ideas of others,how they relate those ideas to their understandingand how they explain their own point of view. • Disputational dialogue • Cumulative dialogue • Exploratory dialogue
Analysing synchronous chatExtensions to: Ferguson, R. and Buckingham Shum, S. (2011). Learning Analytics to Identify Exploratory Dialogue withinSynchronous Text Chat. Proc. 1st International Conference Learning Analytics & Knowledge. Feb. 27-Mar 1, 2011,Banff. ACM Press. Eprint: http://oro.open.ac.uk/28955
Identifying indicators Category Indicator Challenge But if, have to respond, my view Critique However, I’m not sure, maybe Discussion of Have you read, more links resources Evaluation Good example, good point Explanation Means that, our goals Explicit reasoning Next step, relates to, that’s why Justification I mean, we learned, we observed Reflections of Agree, here is another, makes the perspectives of others point, take your point, your view
Cohere.open.ac.ukWeave webs of meaningful connections between ideas
Kmi’s Cohere Rebecca is playing the role of broker, connecting 2 peers’ contributions in meaningful ways web deliberation platform enabling semantic social network and discourse network analyticsDe Liddo, A., Buckingham Shum, S., Quinto, I., Bachler, M. and Cannavacciuolo, L. Discourse-centric learning analytics. 1stInternational Conference on Learning Analytics & Knowledge (Banff, 27 Mar-1 Apr, 2011) http://oro.open.ac.uk/25829
Evidence-hub.net Simon Buckingham Shum, Anna De Liddo & Michelle Bachler
Videoconferencing analytics Flashmeeting video conference: spoken foreign language tutorials Mentor 1 Mentor 2 AV Chat AV Chat 1Session 2 3
Socialized analytics: potential uses Disposition analytics Dispositions can be used to render visible the complex mixture of experience, motivation and intelligences that make up an individual’s capacity for lifelong learning and influence responses to learning opportunitiesBuckingham Shum, S., & Deakin Crick , R. (2012). Learning dispositions andtransferable competencies: pedagogy, modelling, and learning analytics. Paperpresented at the 2nd International Conference on Learning Analytics & Knowledge.
ELLI Profile Web questionnaire 72 items (children and adult versions: used in schools, universities and workplace)
Dimensions of learning powerCritical curiosityMeaning makingCreativityResilienceStrategic awarenessLearning relationshipsChanging and learning
A ‘visual learning analytic’ Basis for a mentored discussion on how learner see themselves, and strategies for strengthening the profile
Connecting withlearner identity SIngleton High School (ratified by the Wonnaruah elders) willy wagtail: didijiri emu: kungkurung snake: ta nipa tang eagle: ka-wul echidna: kuntji kukan platypus: pikan ants: yunrring
Primary school bloggers Creativity Meaning making Strategic awarenessEnquiryBlogger: blogging for Learning Power & Authentic Enquiryhttp://learningemergence.net/2012/06/20/enquiryblogger-for-learning-power-authentic-enquiry
New possibilitiesELLI works from whatlearners say they do Now we can observe what they actually do…
Could a platform generate an ELLI profile from user traces? Different social Questioning and network patterns in challenging may different contexts load onto Critical may load onto Curiosity Learning Relationships Repeated Sharing relevant attempts to pass resources from an online test other contexts may may load onto load onto Meaning Resilience MakingShaofu Huang: Prototyping Learning Power Modelling in SocialLearnhttp://www.open.ac.uk/blogs/SocialLearnResearch/2012/06/20/social-learning-analytics-symposium
Possible future implementationDisposition analytics
Socialized analytics: potential usesContent analyticsVarious automated methods used to examine, indexand filter online media assets for learners.These analytics may be used to providerecommendations of resources tailored to the needsof an individual or a group of learners.
Socialized analytics: potential usesContext analyticsAnalytic tools that expose, make use of or seek tounderstand learners’ contexts. These analytics maybe used alone, or may be employed as higher-leveltools, pulling together data produced by otheranalytics. Context as a dynamic process – a mobile device can present content, options and resources that support learning activities in this location at this time.
Virtual field trip InventoryMandatory OptionalHiking boots Waterproof Project led by Shailey MinochaHand lens clothingGrain-size Sun hatchart Sunscreen Sunglasses Compass 13ºC
Virtual field trip • Do learners engage with all the tasks? • What are the differences between novice and expert practice?
Implementing analytics Customisable dashboards for learners and educators
Implementing analyticsCan we achieve this?•Aligned with clear aims•Huge and sustained effort•Agreed proxies for learning•Clear and standardised visualisation•Driving behaviour at every levelCan we avoid this? Learning with others•Instructivist approach•Stressed, unhappy learners•Analytics with little value for learners or teachers•Omission of key areas, such as collaboration
For more from the OU, seehttp://www.open.ac.uk/blogs/innovating/