I just came back from the 2019 Learning Analytics and Knowledge (LAK) conference hosted by Arizona State University. This conference, surely one of my favorites, is inching to its 10th anniversary next year. I still remember when I was in the 2nd year of my PhD in 2010, Chris Teplovs told me about an exciting, upcoming gathering of 60~ people in Banff, Canada to discuss a thing called “learning analytics.”
After nearly a decade of development, we see clear progress this community has accomplished. The growth of the conference size itself, from 60 to 500+ attendees, while impressive, does not speak to the international sensation around this topic. As someone who first attended LAK as a doctoral student in 2014 and now work on a tenure-track position to grow this area at my institution, I can speak to my lived experiences of observing the rise of local interests in this topic (e.g., the “R1” university I’m part of, the state’s Department of Education), as well as an international passion for learning analytics when I travel to Singapore, Manila, and Beijing.
The rise of learning analytics (the field), LAK (its conference) and SoLAR (its society), coincides with many other things. In education, there is a push for “evidence-based practice” and “data use” in schools (however we mean by evidence or data). In EdTech, words like A.I., adaptive learning, personalization are mentioned in board rooms, websites of philanthropist foundations, and news headlines. Of course, an earlier surge of MOOCs and an ongoing rise of online, lifelong learning demand analytic supports while also offering fertile grounds for learning analytics to grow. In many cases, learning analytics are truly needed and making impossible things possible. But in others, it’s scooped up to be a tag on a product to drive sales.
Realities, for better or worse, just like many other fields we can think of.
First of all, many thanks to the amazing organizing team who has done a wonderful job to make everything real and enjoyable! The program is cohesively organized; conference amenities are thoughtfully arranged; student volunteers are kind and amazingly helpful. These are on top of myself escaping -30C weather in Minnesota for the 20C dessert temperature.
As the community grows bigger, I always found multiple concurrent sessions/talks I hoped to check out. So my impression and report of this conference are surely biased.
I went to this workshops mainly for jokes among Canadians, Aussies, and Serbians :). I was sorted of disappointed because Vita was clearly jet-lagged.
Jokes aside, I’ve been a big fan of Connectivism, George Siemens, and Stephen Downes for years. Many ideas under Connectivism resonate with me, as a big part of my research is grounded in the Knowledge-Building Community model created by Scardamalia and Bereiter.
George kicked off the workshop by briefly explaining its 15-year history. Connectivism emerged in the early stage of the Read/Write web (Web 2.0) and grew out of a dissatisfaction with existing frameworks of learning that were not very helpful in explaining the construction of “modern” knowledge in a networked world.
The core arguments made by Connectivism are:
- Knowledge is networked. This idea is in line with works from Rosen (2012), combinatorial creativity, Nielsen (2013), and Borner (network of science).
- Learning is the process of developing those networks.
- These two ideas apply beyond learning.
At this workshop, George challenged us to work on three domains where learning is networked: Neural; Cognitive/Conceptual; and Social/External. Future work on Connectivism should try to deliver quality empirical research about networks of learning at these levels, detail how things fit together, integrate networks across levels, and articulate human and artificial cognition.
Critiques of Connectivism as a learning theory were mentioned for multiple times. It’s probably accurate to say George found these critiques fair and highlighted that Connectivism meant to provide a framework (or paradigm) to think about knowledge and learning in a fresh way. There is a vast empty space left by Connectivism, including pedagogy, learning analytics, etc.
Stephen Downes joined the workshop remotely and shared a very informative review* of recent work on connectivism. His review looked at scholars’ various interpretations of connectivism, criticisms, connectivism as pedagogy (e.g., Wang, Anderson, & Chen 2018), connectivism as a theory of learning, and properties of a successful network. He offered a number of future directions, including critical thinking and deep learning (not the ML sense), rhizomatic learning, social networking, and cognitive cities. Listening to such a review from one of the creators was helpful.
(*: As a strong advocate of open scholarship, Stephen may have chosen to focus this review on open-access publications.)
Coming into this workshop, I was looking for two things:
- Conceptually, I was wondering why many research papers on cMOOCs used frameworks such as the Community of Inquiry to examine connectivist learning instead of connectivism itself. I was curious about further articulation of connectivism.
- In terms of design – of both connectivist learning environments and analytics – I was looking for design principles and design ideas to inform some of our ongoing projects.
Four principles of connectivism Stephen laid out later was quite useful. These principles include:
- Autonomy of learners
- Diversity of learners, ideas, information, etc.
- Openness of the environment and process.
- Interactivity within the network.
I found these principles valuable even though more work is needed to further unpack and articulate them. Stephen predicted a shift is going to happen (if not yet) even if it’s not necessarily tagged with connectivism.
At the workshop, there were several interesting research talks as well, including analyzing group communication (Nia Dowell), assessing idea development using epidemiological modeling (Sasank Peri), and modeling network dynamics (Renzhe Yu). They represented promising directions to further advance connectivism, especially in the areas of analyzing connective learning processes and developing automated indices about connective learning. Srećko Joksimović also presented findings from an ongoing bibliometric study of the connectivism literature. More details can be found on the workshop website.
I very much appreciated the multi-level network view laid out by George and his urge for us to connect levels. These ideas are informative for my recent thinking, tinkering, and design (since 2017) to:
- Represent, describe, and model learning (at multiple levels) as complex networks
- Design tools for personal, connectivist learning that give learners agency to maintain their networks
- Brainstorm an analytical infrastructure that maintains complexity of all sorts of learning processes
Attending this workshop was a treat!
It was an honor to give an invited talk about Value-Sensitive Design in learning analytics at the FairLAK workshop organized by Kenneth Holstein and Shayan Doroudi.
It is a timely workshop that deals with a topic that’s critical for our work in this community and beyond. I didn’t know there are 21 fairness definitions (Narayanan, 2018; Youtube video). During my limited time spent at the workshop, it is refreshing to see multiple approaches to considering fairness, ranging from creating new metric based on slice analysis to evaluate fairness in predictive models (Josh Gardner), to applying the Socio-Technical Integration Research approach to design processes.
The program and slides of the FairLAK workshop can be found on its website.
Stephanie Teasley and John Stamper asked me to stop by another workshop about curriculum development in learning analytics. It was half-way through after coming from the FairLAK workshop.
Charles Lang and Leah Macfadyen showed us their informative reviews of degree programs related to learning analytics. John Stamper showed us CMU’s Master’s Educational Technology and Applied Learning Science program that interestingly integrates courses from different disciplines and connects deeply with the industry.
Following Charles’ lead, we even got a chance to co-design a class activity to engage students to consider and experience themselves as data producers, data analysts, and analytics consumers. I believe nurturing criticality in all course of a learning analytics program is critical. Interesting activities like this could be very helpful.
We discussed collectively mapping out a learning analytics curriculum. Workshop leaders will be working on compiling a list of courses and syllabi to be shared on the SoLAR website. I look forward to being part of this important work.
There were many interesting workshops happening at LAK.
The Writing Analytics workshop has generously shared their notes that I look forward to dig into. Mladen Rakovic from SFU channeled some “knowledge telling vs. knowledge transforming” wisdom (Bereiter & Scardamalia) I’m familiar with.
The Doctoral Consortium seemed to be going strong as well! I chatted with a couple of student participants during the conference. I’m sure they received great advice from this panel (see below). As you may find out from an upcoming SoLAR 10-year anniversary video, I feel in indebted to this community especially because it has been so invested in emerging scholars. Way to go!
#LAK19 Doctoral Consortium careers panel with special guests @aekrumm @sasha @whitmer joining the co-chairs to share the pros and cons of future trajectories After The PhD...! pic.twitter.com/RheoO3IOH2— Simon Buckingham Shum (@sbuckshum) March 5, 2019
Continue to read Part 2 of the conference report.
- Posted on:
- March 8, 2019
- 8 minute read, 1552 words
- See Also: