Stephen Downes

According to this article, " Data can tell us ‘what’, but not ‘why’. This is the message presented by Prof Bart Rienties, professor of learning analytics at The Open University (OU)." I think he's presenting a narrow view of 'data' as (say) the statistics we get from on learning analytics. Data tells us not only 'what', but also 'why', 'when', 'how much', and 'who cares'. But it's all data - the ... [Direct Link]

Michael Feldstein is participating in the development of "a new courseware platform" and reaping the benefits of deeper insight into learning and development as a result. The layer between content and analytics, he writes, is meaning - software develoeprs call it the 'information model', but my own experience is that this is only half the story; that's the structure, but there's also what logicians would call the 'interpretation' of the data. That's ... [Direct Link]

This is just a short article and mostly a promo for blog posts to come, but it does contain a point that should be highlighted (indeed, Aaron Silvers puts it in all caps in the post): "counts are not analytics". Sure, we need to count things - numbers of hours, how many replies, h-index, and the like. The analysis depends on choosing the right metrics, understanding what story they tell ("All models are wrong, but some are useful."), and comprehending what follows by ... [Direct Link]

The very first sentence below the headline made my eyes roll and things didn't get much better through the duratcion of this "gosh, aren't we great!" article. The 'lessons' are (quoted): start small, use evidence-based research to help share the narrative, and celebrate your successes. I can say that they probably didn't need a six-year analytics prrogram to learn those lessons. The article is mostly a prop for a pseudo-award given to OU last September by dataIQ, &... [Direct Link]

The core argument in this paper (19 page PDF) is that students' perception of feedback from learning analytics has an impact on how they regulate their own learning (see the diagram), which means that (in theory, at least) if you can tailor the feedback to individuals you can have a greater impact on processes like goal-setting. As always (from my perspective) it's really hard to work from such vague terms and imprecise measurements such as, say, self-perception of motivation, to ... [Direct Link]

According to the author, learning analytics "is seldom used to support collaborative learning particularly in face-to-face (F2F) learning contexts," something this paper hopes to correct. Specifically, the analytics in question here provide support for collaborative argumentation. What that means is that a group of people works together in order to jointly create an argument, with each person contributing their own perspectives and ideas. The analytics engine in this case would do ... [Direct Link]

Matt Crosslin writes, "different choices could be made to not create and collect data if the people in control wanted it that way. It is not inevitable that data has to be generated and collected." This is a good point. None of the current capacity to track and record data happened by accident. He continues, "I think that the basic fundamental lens or mindset or whatever you want to call it for publishing research or presenting at conferences about anything from Learning ... [Direct Link]

I saw an item in my feed reader from  Mark Pegrum titled 'Attentional Literacy As A New Literacy' in CJLT, but it just resulted in a link error (so I assume it was from a forthcoming issue? here is the link - good luck). In view of the item on 'feedback literacy' - another new type of literacy - I thought I'd note it anyway. I searched on Google and found a lonely article by Christy I. Wenger with just 1 mention - but it was behind a paywall, so no link here either. I ... [Direct Link]

What behaviours would we take as evidence that students are cheating? This article detects groups of students working together though the similarity of their answers and the proximity in time that these answers were submitted. It then applies analytics to their online behaviour to see if any other sorts of analytics can catch the unauthorized collaborations. They get good grades, but that doesn't set them apart. Nor do their interactions with the online course. Nor does the timing of ... [Direct Link]

This article introduces the Learning Analytics and Visual Analytics (LAVA) model "as a conceptual framework through which VA can be seamlessly integrated into the LA process." The idea is to extend trhe human in the loop approach used in LA to include VA as well. "The concept of VA," write the authors, "can be better understood based on the sense-making loop for visual analytics... A visualization is presented to the user based on the initial analysis of the dataset... ... [Direct Link]

This excellent report (55 page PDF) will introduce you to the Connected Learning Analytics (CLA) toolkit. This toolkit extends learning analytics beyond the LMS, and specifically, to a set of six social media platforms (Twitter, Facebook, Wordpress, GitHub, Trello, Slack). It is based on Kitto's work from 2018 developing the Connected Learning (CL) Recipe, released as open source on GitHub. The data is saved in a learning record store (LRS) using xAPI. It is converted yet again using a ... [Direct Link]

This is a pretty good article (20 page PDF) that looks at the limits of analytics in designing learning programs and makes some recommendations for higher-level constraints. "The network structure does not reveal everything," writes Daniel Domínguez Figaredo. Things as varied as pedagogical approach and harmful content need to be considered and regulated from a wider context. "In many cases, researchers using social network analysis take into consideration the ... [Direct Link]

This code builds on Jisc's code of practice for learning analytics and extends the discussion to include "additional issues raised by the use of data for wellbeing purposes... on the use of data in delivering wellbeing and mental health support." I think this is a bit of a departure for Jisc, as we would normally think of health and well-being as something health professionals deliver, not educators, though while acknowledging this the code says "the day-to-day operation of ... [Direct Link]

I read this document (24 page PDF) as part of my research into ethics for learning analytics (and I note, in passing, the almost complete absence of references to ethics for teachers among the discussions of learning analytics). What caught my interest in this document was the progressive and wide-reacing account of professional development it provides; it's far more than mere classroom training, and includes a wide degree of involvement in the community as part of its desciption of ... [Direct Link]

The authors write, "this survey will be especially helpful for those with a basicunderstanding of machine learning, interested in (i) getting a comprehensive but accessible overview of many fundamental deep learning concepts and (ii) references and guidance in helping ramp up implementation." I can't speak to (ii) but I can say that the explanations are clear and concise, and that the paper as a whole will help a lot of people organize their thinking about deep learning. Topics ... [Direct Link]

This article (6 page PDF) draws explicity the similarity between the spread of a disease and the spread of an idea, and measures that spread in Twitter conversations related to several recent conferences in learning analytics.  "instead of focusing on individual knowledge acquisition," the authors write, "we argue for attending to the social process of idea expression, spread, and evolution." To this end, they write, "the Lifespan of an Idea framework is proposed.&... [Direct Link]

Colin Beer argues - correctly - that "techno-centric approaches to learning analytics, where the focus of attention is on historical data, are fundamentally flawed. I’m not sure we truly understand this yet as a sector." The reason for this is neatly illustrated by the three-body problem. "If we have a class with only three students, we cannot begin to perceive or quantify the variables that will contribute to their success or otherwise." Students and learning are ... [Direct Link]

I had the plan of turning a talk I gave last year on Ethics, Analytics and the Duty of Care into an article. I've been working on it for a while, and it has gradually become a book project (my raw raw notes are here and my draft text is here). As I suggested in Brazil when I gave the talk is that there is no simple agreement on ethics, and issues that are presented as clear-cut (such as: we're all against surveillance!) are far from obvious when we actually look at them. Anyhow, this ... [Direct Link]

This is a pretty good paper. The authors draw on three meta-analyses of learning analytics literature in order to create a framework for describing learning analytics projects. They then apply this framework to four use cases: socal learning analytoics for prediction and for reflection, and individual learning analytics for the same two functions. I think this reading of learning analytics is a bit narrow, and the divisoion between social and individual a bit arbitrary, but the tables ... [Direct Link]