Analyzing Learning FLows in Digital Learning Ecosystems

Analyzing Learning Flows in Digital Learning EcosystemsMaka Eradze, Kai Pata, and Mart Laanpere :: Tallinn University, Estonia

Socio-technical transitions

Geels 2002

Mobile communication generations

Moodle servers: global stats

Three generations of TEL systems

Dimension 1.generation 2.generation 3.generation

Software architecture

Educational software Course management systems

Digital Learning Ecosystems

Pedagogical foundation

Bihaviorism Cognitivism Knowledge building, connectivism

Content management

Integrated with code Learning Objects, content packages

Mash-up, remixed, user-generated

Dominant affordances

E-textbook, drill & practice, tests

Sharing LO’s, forum discussions, quiz

Reflections, collab. production, design

Access Computer lab in school

Home computer Everywhere – thanks to mobile devices

Analytics Only feedback for learner

Frequency-based usage statistics

Interaction & uptake analytics

Digital Learning Ecosystem

Ecosystem (biol.) is a community of living organisms (plants, animals and microbes) in conjunction with the nonliving components of their environment (e.g. air, water, light and soil), interacting as a system.

DLE is an adaptive socio-technical system consisting of mutually interacting digital agents (tools, services, content used in learning process) and communities of users (learners, facilitators, trainers, developers) together with their social, economical and cultural environment.

Dippler: a prototype DLE

Learning interactions

Wagner (1994): reciprocal events that require at least two objects and two actions. Interactions occur when these objects and events mutually influence each other

Dyadic model of learning interactions (Moore, 1998): learner-learner, learner-teacher and learner-content

Equivalence theorem by Anderson & Garrison (1998): reduction in one dyad can be compensated by increase in another

Suthers (2011): interaction is fundamentally relational, so the most important unit of analysis is not isolated acts, but rather relationships between acts

Learning analytics: a critical view

Most of the LA research is conducted in closed LMS context using frequency-based statistical analysis

Only learner-content interactions are studied, not relations between the interactions

Social Network Analysis (SNA) is focusing on teacher-learner and learner-learner interactions, but neglects the aspects of quality and dynamics in interactions

Communities of Inquiry (CoI) approach focuses on quality and dynamics of learning interactions, but it is not scalable

Sequential analysis of learning flows

In addition to frequency-based statistics, exploratory sequential data analysis is needed for analytics of learning flows in DLE

In Dippler: extending Activity Streams (activitystrea.ms) vocabulary with pedagogical Action Verbs and Objects

TinCan API or xAPI: specification for learning technology that makes it possible to collect data in a consistent format about the wide range of experiences a person has (online and offline)

Uptake Framework (Suthers & Rosen, 2011): Uptake happens when a participant takes aspects of prior events as having relevance for ongoing activity; UF results with contingency graphs that can visualise media dependency, temporal proximity, spatial organization, semantic relatedness, inscriptional similarity

Sample scenario

Collaborative concept mapping: identifying the core set of concepts for a given domain along with and their relations with each other, using a digital concept mapping tool

Task is connected with some key concepts in domain ontology and also with a specific learning outcome

Event transcript in activity stream: In Assignment 3, John adds a relation to conceptmap12 with CMapTool at 12:30 12-07-13.

Results: contingency maps and uptake diagrams are created and fed back to learner and teacher, irregular patterns notified

Conclusions and future research

Combining xAPI with Uptake Framework creates new opportunities for Learning Analytics and has several advantages: Recording interactions in dyadic events will encompass the

processes, traces, domains; feedback loop for teachers & learners The relations with a domain will be identified and generalised

through semantic annotation of events and artifacts Enables recording of the interactions that take place in distributed

and partly user-defined digital ecosystem Advanced learning interaction analytics is automated and scalable

Next steps: building xAPI Learning Record Store for Dippler and extending it to wider ecosystem, also to the physical world