Skip to content

LINK Research Lab: Fall Speaker Series

At LINK Research Lab, we have a full slate of speakers for fall, including topics on distributed learning, synchronous instruction, success for under represented students, learning analytics, engagement, design based research, openness, flipped classrooms, health and the built environment, mentorship, and wonder. The full schedule is here. We will be streaming the events online and are exploring options for asynchronous interaction as well. If you’d like to be informed of event details, recordings, and links to live sessions, please register.

Bundling and Re-bundling

I’m at the Knewton Symposium – an event focusing on the future of digital learning. This is the second year that I’ve attended. It’s a small event (last year had ~20 attendees, this year it’s closer to 60+). Knewton brings in a range of speakers and leaders in education, ranging from startups to big edtech companies and publishers to faculty and advocates for some type of change. The conversations are diverse, as can be expected when publishers and open education advocates as well as VC firms and academics share the same stage.

The narrative of educational change is more stable than it was even a few years ago and it’s reflected in this symposium. In 2011, everything was up in the air: universities were dead, faculty would be replaced by MOOCs, California would solve its education crisis by partnering with a small startup, and so on. Now the narrative has coalesced around: 1. economics and funding, 2. access and affordability, 3. innovation and creativity, 4. data and analytics, 5. future university models. While I’m interested in all five of those narratives, particularly the way in which these are being framed by university leaders, vendors and startups, and politicians, I’d like to focus here on one aspect of the conversation around future university models: unbundling.

Unbundling is an appealing concept to change mongers. The lessons of the album and mp3′s is strong with these folks. MP3s lead to newspapers which lead to music and media in general. Since change mongers (a species native to Silicon Valley but now becoming an invasive species in numerous regions around the world. Frankenfish comes to mind) do not have much regard for nuance and detail, opting instead for blunt mono-narratives, unbundling is a perfect concept to articulate needed change.

There are a few things wrong with the idea of unbundling in education:

1. Unbundling is different in social systems than it is in a content only system. An album can be unbundled without much loss. Sure, albums like The Wall don’t unbundle well, but those are exceptions. Unbundling a social system has ripple effects that cannot always be anticipated. The parts of a social system are less than the whole of a social system. Unbundling, while possible in higher education, is not a zero sum game. The pieces on the board that get rearranged will have a real impact on learners, society, and universities.

2. When unbundling happens, it is only temporary. Unbundling leads to rebundling. And digital rebundling results in less players and less competition. What unbundling represents then is a power shift. Universities are today an integrated network of products and services. Many universities have started to work with partners like Pearson (ASU is among the most prominent) to expand capacity that is not evident in their existing system.
Rebundling is what happens when the pieces that are created as a sector moves online become reintegrated into a new network model. It is most fundamentally a power shift. The current integrated higher education system is being pulled apart by a range of companies and startups. Currently the university is in the drivers seat. Eventually, the unbundled pieces will be integrated into a new network model that has a new power structure. For entrepreneurs, the goal appears to be to become part of a small number of big winners like Netflix or Google. When Sebastian Thrun stated that Udacity would be one of only 10 universities in the future, he was exhibiting the mentality that has existed in other sectors that have unbundled. Unbundling is not the real story: the real issue is the rebundling and how power structures are re-architected. Going forward, rebundling will remove the university from the drivers seat and place the control into the re-integrated networks.

Congrats to Paul-Olivier Dehaye: MassiveTeaching

In a previous post, I commented on the Massive Teaching course at Coursera and that something odd was happening. Either Coursera deleted the prof from the course or the prof was running some type of experiment. It now appears to be primarily the latter.

The story has now been covered by The Chronicle (here and here) and Inside Higher Ed (here). Thoughtful reflections have been provided by Rolin Moe and Jonathan Rees. Participants on Twitter have also had their say. The general consensus is that “wow, this is weird”. Coursera has deftly pushed everything back to the University of Zurich, who in turn has pushed it onto the prof, Paul-Olivier Dehaye. Commenters have been rather cruel (I know, shocking to have mean people on the internet), going so far as to question Dehaye’s sanity. OT: Favourite comment of the day: “Moocs are demonic, and unhuman.”

There is plenty of blame to go around. Dehaye has not publicly commented. Coursera very quickly washed its hands of the situation. What Dehaye did was inappropriate and might have crossed a few ethical boundaries. That’s an important angle, but not one that I want to pursue here. Three substantial concerns exist:

1. Coursera has been revealed as a house of cards in terms of governance and procedures for dealing with unusual situations. While Coursera promotes itself as a platform, something that I wrote about a few years ago, it is more Frankensteinian than functional. MOOCs were developed so quickly and with such breathless optimism that the architects didn’t pay much attention to boring stuff like foundations and plumbing. What is the governance model at Coursera? Is there anything like a due process to resolve conflicts? And a range of questions around content ownership and learner data.

I have a colleague who taught on Coursera recently. He was unable to get access to data that had previously been promised. In a university, there is a counterbalancing process to these types of conflict or disagreements. At MOOC providers, the company rules. This is fine at Facebook, but Coursera is essentially a leech on the education system – getting teaching for free while exploring new ways to monetize the process. (Wait. Doesn’t that make them the Elsevier of teaching and learning? Content and teaching free. Monetize the backend.)

My point here is that the governance structure that underpins university is lacking in MOOC providers. It is not a balanced and equitable system. There are many fissures in the MOOC model and as providers become more prominent in education these fissures will become more evident. If companies like Coursera and edX expect to be able to make decisions on behalf of faculty and partner universities, conflict is inevitable. A transparent process is required.

2. University of Zurich has an obligation and responsibility to its faculty. Where a university’s reputation and identity can be launched internationally in a MOOC, leadership should have some quality control process in place. Is the university so poorly informed about online learning that simply giving a faculty member keys to the kingdom without some guidance and direction was assumed to be a good approach? There is much blame to be shared and it should fall in the following order: 1. Coursera, 2. U of Zurich, 3. Dehaye

3. Criticism ranging from a poorly designed course to poor ethics has been directed to Paul-Olivier Dehaye. Most of it is unfair. There have been some calls for U of Zurich to discipline the prof. Like others, I’ve criticized his deception research and his silence since the course was shut down. Several days before the media coverage, Dehaye provided the following comments on his experiment:

“MOOCs can be used to enhance privacy, or really destroy it,” Dehaye wrote. “I want to fight scientifically for the idea, yet teach, and I have signed contracts, which no one asks me about…. I am in a bind. Who do I tell about my project? My students? But this idea of the #FacebookExperiment is in itself dangerous, very dangerous. People react to it and express more emotions, which can be further mined.”
The goal of his experiment, Dehaye wrote, was to “confuse everyone, including the university, [C]oursera, the Twitter world, as many journalists as I can, and the course participants. The goal being to attract publicity…. I want to show how [C]oursera tracks you.”

There it is. His intent was to draw attention to Coursera policies and practices around data. Congrats, Paul-Olivier. Mission accomplished.

He is doing exactly what academics should do: perturb people to states of awareness. Hundreds, likely thousands, of faculty have taught MOOCs, often having to toe the line of terms and conditions set by an organization that doesn’t share the ideals, community, and egalitarianism that define universities (you can include me in that list).

The MOOC Mystery was about an academic doing what we expect and need academics to do. Unfortunately it was poorly executed and not properly communicated so the message has been largely lost. Regardless, Dehaye has started a conversation, raised a real concern, pushed buttons, and put a spotlight on unfair or opaque practices by organizations who are growing in influence in education. Yes, there are ethical concerns that need to be addressed. But let’s not use those ethical concerns to silence an important concern or isolate a needed narrative around what MOOCs are, how they are impacting higher education and faculty, and how control is being wrested from the people who are vital counter-balancing agents in society’s power structure.

Paul-Olivier – thanks. Let’s have more of this.

I was wrong

I’ve made statements late last year to the effect that “corporate MOOCs will be the big trend in 2014″. I was wrong.

Recently, with CorpU and Reda Sadki, I ran an open online conference on corporate MOOCs. We put together a strong line up of presenters and topics and I expected reasonably strong turnout as the topic was timely. While we had a large number of signups, we only had 15-30 people attend each session. The sessions were generally one-way information flow (from the presenter). Attendees appeared to be reluctant to share experiences and views. I’m not sure if this was due to corporate interests in preserving and not sharing information or if we just didn’t hit on the right topics.

The recordings of most sessions are available here (we had a few requests to not record sessions by presenters). Some excellent presentations!

Aside from not having the engagement I was hoping for, I was interested in several points raised during the event:
- Corporate MOOC completion rates are in the 70-80% range
- Coursera is heavily focused on providing branded “turn key” content for corporation training
- Systems like WorldBank are developing MOOCs as an integrated part of their overall online or digital learning strategy
- Several corporations, notably Google and SAP, are deep in the rabbit hole of MOOCs already and are reporting position experiences for both employees and customers who have taken their courses
- Consulting services such as Parthenon are deeply engaged in MOOCs and helping organizations plan for and deploy them.
- The costs of MOOCs are significant in terms of capital and time and effort of people. It’s not as simple a process as many assume when they start.
- Military organizations are exploring MOOCs and alternative teaching/learning approaches and are reporting promising early results. But we can’t tell you everything. It will be declassified in 2050.
- Organizations are primarily using MOOCs for internal learning, marketing, connecting with customers, and “teaching” suppliers.

Something weird is happening at Coursera

Something weird is happening at Coursera. I’m not sure what it is exactly, but have boiled it down to two options. Both are problematic.

A bit of background

About two months ago, I posted a short article on a DesignJam that we hosted at UT Arlington. The designjam brought together numerous folks who had some interest in teaching and learning online, often at a massive scale (i.e. MOOCs). Paul Olivier Dehaye commented on the post and described his interest in running a dual-track MOOC, blending instructivist and more collaborative. He was referring to the Massive Teaching course on Coursera that he was to run in June. I’ve been continuing to refine my thinking on this since the designjam, but I had not been following Paul’s course. Today, Apostolos Koutropoulos posted about social experiments and confusion at Coursera. I did a bit of backtracking on Paul’s tweet stream.

and

and finally, in response to a tweet asking Paul what was happening, he replied

Two options:

1. Coursera has removed a faculty member from a course for some reason without explanation
2. Paul is running a fairly elaborate social experiment

I am uncomfortable with both. If Coursera has removed the course or the faculty member, some explanation is required, both for the sake of the faculty member and the student. The transparency of MOOC providers is rather poor. If Facebook randomly deleted people, it would cause angst. Coursera doesn’t state the conditions under which a faculty member can be removed or a course cancelled. Universities and faculty spend enormous time and resources developing and running courses. Students devote significant hours as well. Everyone deserves an explanation.

If Paul is running an experiment, well, that raises a range of ethical issues around active experimentation with learners. Kate Bowles links to paper and a Google doc that raises additional questions. Given heightened concerns about ethics in social media and experimentation on users, MOOC providers and faculty need to be clear on any research and analytics being conducted.

Activating Latent Knowledge Capacity

Last week, we wrapped up another successful Learning Analytics Summer Institute at Harvard. The recordings of most of the talks and panels can be found here. Since we were already in town, Dragan Gasevic and I were invited by edX to give a talk to their staff and member institutions (we are running a course on edX in fall on Data, Analytics, & Learning).

The focus for the talk at edX, slides below, was to try and get at what is wrong with MOOCs and education in general. To answer the challenge of “what is wrong with education” it’s helpful to step back a bit and consider two challenges.

1. We aren’t connecting

Historically, society has created knowledge institutions that mirror what is done with information in a particular era – see McNeely & Wolverton. In this line of reasoning, we can best understand the future of education by understanding what is being done with information today. After about a decade of experience with web 2.0, social media, participative technologies, it’s not unreasonable to state that at least a segment of society today recognizes information as multi-authored, participative, distributed, and networked.

In education, many of us have been advocating for networked learning (or as Stephen Downes and I have been articulating it, connectivism). Academic conferences and even the K-12 space has turned to networks as a way to describe what learning is and how it happens. The one draw back to networked learning is that while we have managed to advance conversation on the fragmentation of learning so that it is not a cohesive whole created solely by the instructor, we have not yet advanced the process of centring or stitching together fragmented parts into cohesive wholes for individuals. Some rudimentary progress includes the use of #hashtags to stitch together distributed conversations but this only provides a one medium aggregation. The best implemented model for pulling together multi-platform conversations that I’ve seen to date is Downes’ gRSShopper. That leaves us at a difficult point educationally. Progress has been made on pulling centralized information elements apart (this is particularly evident in media with newspapers or TV news programs – I get the majority of my news in bits and pieces through a mess of different tools and sites), but we haven’t yet developed the technologies that will allow pulling things back together into coherent, personally owned, wholes.

This is no small challenge. In many ways, this is where computing was in two separate phases: pre-Microsoft Office and pre-Facebook. I remember when I used to work with distinct software tools like Quattro and WordPerfect (before they were owned by one company). Moving data between different software was a pain. MS came along and blessed society with Office – an integrated suite. It pulled together what I used to do in several different tools. Facebook plays a similar integrative role for participative technologies. For people who had been blogging since late 1990′s or early 2000′s, Facebook wasn’t of much value. Between flickr, del.icio.us, blogs, RSS readers, and wikis, we were living the distributed, networked, learning dream. Unfortunately, only a small percentage of society wants to deal with a range of 10 different tools. Ease of use and low-barrier to entry rules the day. Facebook allowed anyone to start sharing images, ideas, and form social networks and to do so in a single tool with similar functionality across different activities. My social network used to consist of the people in my RSS reader. Facebook made connecting easy and they were rewarded as a billion+ people joined. The key lesson here is that integrative technologies, in spite of the current app trend, draw greater numbers of users than single functionality tools.

The importance of integrative toolsets for learners cannot be overlooked. It is unreasonable to expect a learner to care about the same issues that an instructor of a participatory course cares about. While concerns of access, participation, and equity might be important to me, a learner may well enter a course with the primary goal of learning a skill or concept. My values may not be her values.

2. Latent Capacity

Technology cannot be reduced to a single narrative or outcome. While “web 2.0″, as a term, symbolized participation and collaboration, it is really a multi-narrative strand where some people were enabled and others were shut out, some were given a platform and others lost a platform, some connected with their readers/fans and others were exposed as [insert label] to their fans/readers. There are many narratives to describe the tools that today define how people interact. I have been grappling with understanding the prominent or even dominant impact of technology – i.e. what is one aspect of technology that is most pronounced and most misunderstood? Keeping in mind that a single narrative has shortcomings, I’ll argue that activation of latent capacity is the driving element of every successful technology of the past 15 years. Uber uses latent car capacity. Airbnb, latent physical space capacity. Twitter/Facebook, activate multiple latent capacities: sharing, social connections, and images. The Arab Spring, now sadly turned into a rather harsh winter, and Occupy Wall Street activate the latent power capacity of individuals. A system of control and oppression can be challenged when people take up their power, their voice.

In education exists the most substantive latent capacity in society. A classroom consists of 30 (or sometimes 300) people listening to a teacher teach. The knowledge and creative capacity of any class is stunning. Unfortunately, this knowledge is latent as the system has been architected, much like a dictatorship, to give control to one person. In many cases, students have become so accustomed to being “taught” that they are often unable, at first, to share their knowledge capacity. This is an experience that I have had in every MOOC that I’ve taught. The emphasis in MOOCs that I’ve been involved with is always on learners taking control, learners joining a network, or learners becoming creators. In a Pavolovian sense, many learners find this process disorienting and uninviting. We have been taught, after a decade+ of formal schooling, to behave and act a certain way. When someone encourages a departure from those methods, the first response is confusion, distrust or reluctance.

I’ll call my theory of knowledge and learning “100 people in a room”. If we put 100 people in a room, the latent knowledge capacity of that room in enormous. Everyone in this room has different life experiences, hobbies, interests, and knowledge. We could teach each other math, physics, calculus. We could teach poetry, different languages, and political theory. The knowledge is there, but it is disconnected and latent. Much of that knowledge is latent for two reasons: 1) We don’t know what others know, 2) connections aren’t made because we are not able with our current technologies to enable everyone to speak and be heard.

Personal Knowledge Graph

To address these shortcomings, I’ve been arguing for the development of something like a Personal Knowledge Graph (PKG). The main idea is that learners need a way to express and articulate what they know. This can be done through someone explicitly stating “I know this” or it could be mined or inferred. Learners need to own their PKG but it should be shareable with schools, companies, and peers.

Once we know what people know, we have a chance to activate latent knowledge through social and technological approaches. The work that Dragan Gasevic has done with his doctoral students indicates that learners begin to use hashtags as a cognitive agent. In some cases, a hashtag becomes a more important agent than a faculty member. In other instances, recommender systems could connect individuals who have complimentary and/or opposing knowledge graphs. This leads to new pedagogical models and changing roles for universities, notably a transition from spraying the same content to all learners to a more nuanced (knowledge gap filling?) approach.

Education is approaching where the web was in mid-2000′s – a growing range of technologies providing certain opportunities for learning and interaction, but largely fragmented. Education is waiting for it’s latent capacity activating tool, or at minimum, a means of giving each learner the ability to stitch together a coherent interpretation of a knowledge domain. Of course we need feedback loops and systems of recognition. It is not enough that I state I know something. Peers, faculty, and employers should be able to comment on my claims and I should be able to provide evidence. When I do not understand a concept correctly, there should be processes for correction.

If, when, education begins to focus on activating the knowledge of individuals rather than primarily focusing on single point knowledge pontification, new concerns will arise. For example, how can creativity be encouraged when learners receive personalized content addressing knowledge gaps? What happens to formal assessment? What role does expertise play in a room of 100 knowledgeable people? The transitions underway in society, in knowledge, and in universities, are long term and won’t be played out in the next few years. It’s a decades long transition. But it is important to begin challenging legacy assumptions and start considering, however imperfect our ability to see it today, what an education system looks like when we activate latent capacities of all participants.

Attend (online) the Learning Analytics Summer Institute

Last year, we held the Learning Analytics Summer Institute (LASI) at Stanford. This year we will hold LASI at Harvard. The event starts tomorrow and runs for three days (June 30 – July 2). Our interest and mission with the Society for Learning Analytics Research is to make data and algorithms open and accessible to researchers in order to create transparency around how analytics are being used in teaching and learning. As such, we will be live streaming LASI. The schedule of speakers, and we have an amazing set of panels and keynotes, is available now (scroll down the page for the live video feed). We are also running a global network of LASI-Locals in Hong Kong, Egypt, South Africa, Netherlands, Spain, Latin America, UK, and other regions. If you are interested in learning analytics and how they are being deployed by researchers and students, please join us online. It’s a distributed global conversation with a few thousand peers who are exploring data, analytics, and learning.

Tag for the event: #lasi14

If you blog, please follow directions here on adding to the blog feed (it’s not quite gRSShopper)

Online event: Scaling Corporate Learning

We’ve now posted additional resources for the Scaling Corporate Learning online conference (June 18 & 19):

Schedule

Speaker Bios

If interested in joining, register here

MOOC Research Initiative: Reports Released

I’m pleased to announce that the reports from the MOOC Research Initiative (MRI) are now available. In the spirit of useful science, each grantee was asked to provide a one sentence overview of their research findings.

I enjoyed working on this grant. The traditional method of exploring a knowledge domain takes years. With MOOCs, given their rapid development, this didn’t seem feasible. The approach that we took with MRI was small grants with a rapid review cycle, interim presentation of results through our MRI conference with other grantees in order to increase idea exchange, and short funding cycle. We have numerous papers that will be published in a special issue of IRRODL. Science should look more like this process, especially in fields where things are changing quickly, than the long multi-year cycle that we currently see.

New MOOC: Data, Analytics, & Learning

I’ve run a range of open courses on a fairly broad range of platforms: D2L, Moodle, Instructure, a mess of social media tools, and (most frequently) with Stephen Downes’ gRSShopper.

This fall, together with colleagues, I’ll be offering an open course on edX: Data, Analytics, and Learning. From the description:

In education, the use of data and analytics to improve learning is referred to as learning analytics. Analytics have not yet made the impact on education that they have made in other fields. That’s starting to change. Software companies, researchers, educators, and university leaders recognize the value of data in improving not only teaching and learning, but the entire education sector. In particular, learning analytics enables universities, schools, and corporate training departments to improve the quality of learning and overall competitiveness. Research communities such as the International Educational Data Mining Society (IEDMS) and the Society for Learning Analytics Research (SoLAR) are developing promising models for improving learner success through predictive analytics, machine learning, recommender systems (content and social), network analysis, tracking the development of concepts through social systems, discourse analysis, and intervention and support strategies. The era of data and analytics in learning is just beginning.

I’ll provide more information soon about the design of the course – we are focusing on dual structured and self-organized approaches to the course.