Ed-tech in lifelong learning - a literature review

Free Online Education: Too Much Choice?

In the last decade, the explosion of platforms such as Khan Academy (2008), edX (2012) and Coursera (2012) has led to a proliferation in educational content online; Coursera alone has jumped from 350 courses in 2013 (Mitchell, 2013) to more than 2400 in 2018 (CNBC, 2018). This rise in popularity is possibly a response to the fact that a large proportion of people (74%) in a modern society like the US see themselves as lifelong learners (Pew Research, 2016), and many corporations and non-profits are trying to satisfy this demand.

Lifelong learning and particularly the proliferation of educational content bring their own set of challenges to the table, however. There is an increased responsibility for the learner in planning their learning (Goodyear, 2000), in a world where the large numbers of freshly released courses are piled on an already existing trove of educational content (books, videos, articles, etc.). Indeed, how does the lifelong learner inventorize what they know, and what they don’t know? How do they decide what educational activity might be optimal for them to undertake next? How do they remain committed to their lifelong education goals?

We will focus our research on the motivation of students in online learning, as well as the question of autonomous systems versus human tutoring systems, to conclude with a set of recommendations to guide the design and development of our application.

Motivation in Online Learning

Self-regulated learners develop the fundamental trait of autonomy by identifying needs, setting goals, planning their learning, selecting resources and learning strategies, through practice, and by monitoring and assessing their progress (Reinders, 2010). Additionally, effective self-learners seem to be motivated by intrinsic factors, such as the joy of learning or the usefulness of learning in their lives, rather than extrinsic factors such as grades or completion certificates, while less effective self-learners tend to be the opposite. Furthermore, effective self-learners seem to set their own learning paths, sometimes dropping modules that they are already familiar with, and setting their own orders of completion, while less effective self-learners are more prone to following the established order (Littlejohn et al., 2016).

But how might these insights on self-learner motivation materialize in specific recommendations for application developers such as ourselves? A study of the development of an ed tech application recommends that apps should “complement existing educational platforms”, “provide comprehensive support for students”, “provide different perspectives for information analysis”, as well as “offer the learner goals, standards or comparison criteria to analyse their behavior” (Pérez-Álvarez et al., 2018)

Further recommendations for interaction designers in educational technology are to provide rationales for content and activities, emphasizing why the learning is important and worth doing, enabling learners to make connections between contents and their everyday lives, provide options for communication, both synchronous and asynchronous, and providing care not to overwhelm users with information and additional resources (Hartnett, 2018)

Intelligent Systems in Educational Technology

Despite a recency bias that may lead some to believe that educational technology is a recent trend in academia, the field is in fact decades-old. For example, landmark research in advanced topics such as intelligent tutoring systems dates as far back as the 1970s. Arguably, however, the promise of a delivery of said systems at scale has failed, as they are difficult to implement and fail to leverage fundamental human strengths in tutoring (Baker, 2016).

Delving into some specific difficulties in recommender systems, we find that cold start problems are very pronounced - learners often start from a zero-state, as opposed to say, movie watchers - and that recommendation models are hard to define - popularity might work for movies, but not always for education (Draschler, 2008)

A more interesting approach might therefore be the development of mixed systems that leverage the strengths of human beings as tutors or educators, helped by state-of-the-art data analytic capabilities (Baker, 2016) One such approach is the concept of a learner network, a network that is “populated with many users and learning activities provided by different stakeholders. Each user is allowed to add, edit, delete or rate learning activities at any time”. This system connects “people, institutions, learning artifacts and autonomous agents in such a way that the human network becomes self-organized and will give rise to effective lifelong learning” (Koper, 2003)

Developing a Solution

Lifelong learners need help in planning learning and selecting resources (Reinders, 2010). The temptation to build a fully automated system might be strong, but on the one hand, it seems clear that learners benefit from social interaction (Hartnett, 2018) and that the promise of fully intelligent educational systems at scale is difficult to implement and fails to leverage natural strengths of human beings (Baker, 2016). We should therefore prioritise features that allow the interaction amongst learners and instructors, using intelligent systems to bolster these relationships as needed, along the vision of Koper’s Learning Networks, where learners can create their own learning activities, build their own learning paths, and share their learning activities and their learning plans with peers and institutions.

From a student perspective, effective self-learners are motivated by intrinsic rewards, such as the pleasure of learning or the link of learned knowledge with real-life situations. Similarly, there is evidence that effective self-learners are not completionists - they might not favor completing a resource in order or at all, picking only what interests them (Littlejohn, 2016). This seems to suggest to us that we should aim to have a tool that allows for a great amount of flexibility, whereby changes are permitted on the fly and partial completion of objectives might still be regarded as a victory.

Bibliography

Baker, Ryan S. (2016). International Artificial Intelligence in Education Society.

CNBC (2013). 2018 Disruptor 50 Full Coverage, Coursera.

Drachsler, H., Hummel, H. G. K., & Koper, R. (2008). Personal recommender systems for learners in lifelong learning: requirements, techniques and model. International Journal of Learning Technology, 3(4), 404-423.

Goodyear, Peter (2000). Environments for lifelong learning: ergonomics, architecture and educational design.

Hartnett. Maggie (2018). Motivation in Online Education (Book, Springer)

Koper, E.J.R. and Sloep, P.B. (2003) Learning Networks: Connecting People, Organizations, Autonomous Agents and Learning Resources to Establish the Emergence of Lifelong Learning, Heerlen, The Netherlands: Open University of the Netherlands

Littlejohn, Allison; Hood, Nina; Milligan, Colin; Mustain, Paige (2016). Learning in MOOCs: motivations and self-regulated learning in MOOCs. The Internet and Higher Education, 29.

Perez-Álvarez, Ronald; Maldonado-Mahauad, Jorge, Pérez-Sanagustín, Mar (2018). Journal of Universal Computer Science, 24(8):1090-1109

Pew Research Center (2016). Lifelong Learning and Technology

Reinders, H. (2010). Towards a classroom pedagogy for learner autonomy: A framework of independent language learning skills. Australian Journal of Teacher Education.

Waldrop, M. Mitchell (2013). Massive Open Online Courses, aka MOOCs, Transform Higher Education and Science, Nature Magazine.