dual layer MOOC

Scaffolding an Entire University to Open Learning

A lot of what I have been blogging lately is just me struggling through various ideas surrounding this whole “Dual Layer MOOC” design idea. Probably the whole term “Dual Layer” is a misleading descriptor anyways. Multiple pathways is better, but since that term already has specific designs attached to it, its hard to fight against that. “Multiple pathway” courses still tend to be “multiple siloed pathways” in which five or ten or how many ever specific defined pathways are given. That’s not really the goal that instructors have for this course.

The underlying goal is create a course that emphasizes diversity, experience, and autonomy in learning, to borrow a description from Stephen Downes.  The problem we are dealing with is the reality that the entire University system is set up in an instructivist manner that values all students going through the same path in each course in order to pass the course by doing exactly what they are told. Students are so used to this system that they are comfortable with it and start freaking out if they are forced to take an open course. To borrow a statement from George Siemens: “We can’t force students to be open.”

So the dual layer MOOC is not about blending cMOOCs and xMOOCs as much as creating a scaffold for those students who are used to instructivist learning to dip their toe in and try out networked learning – if they want. But there are those that want connected deconstructed learning from the beginning, so that option has to be a viable one from the beginning also. If at any time we create narrow pathways that force students to scaffold from instructivism to connectivism, we leave diversity, experience, and autonomy behind. So the door has to remain open, but the learner has to choose when to pass through it.

So this is not a case of the xMOOC wagging the cMOOC tail, or vice versa. If it looks that, its just because I am failing to create adequate metaphors to explain what has been coming out of the design meetings. I still like the play dough metaphor best (we’re just throwing a bunch of play dough cans on the table and learners can pick them up and use them as the like in groups or individually or even just leave the room and go get their own play dough) – but that makes for a lousy blog diagram :)

So, in a lot of ways, I just see this dual layer thing as a step on the process of scaffolding the university system from instructivism and teaching to “sharing the process of thought and inference and discovery with those around you” (to quote Stephen Downes again). That sharing process is the main reason why I started blogging so much about the dual layer MOOC – it will change and even possibly go away. I’m just sharing my process openly. And the feedback I have received has been awesome – so it has been a worthwhile process and will continue time permitting.

Learning Design Versus Learner Design in the Dual Layer MOOC

So I want to take a step back and look at a bigger picture for a  moment. The previous diagram I created was meant as more of a “learner design” perspective, in the sense that the dual paths would be the two main possibilities for learners. This was meant to look at the course from the eyes of a learner. But, the diagram breaks down in many ways since this will not be a traditional course. This course will attempt to deconstruct what it is to learn, be a learner, and to move through a course.

There has been a lot of really good feedback on the diagrams. Mike Caufield makes a great suggestion for considering the idea of federation in this course. I’ll have to figure that one out some more and unpack that in another post. I want to zero in on a few big design considerations first.

Stephen Downes asked “But is that really distributed, the way a cMOOC would be? At a certain point, the movement to collect people into single-site courses collides with the movement toward things like indieweb and reclaim your domain.” Good questions, and very true. The main idea behind this dual design is to be distributed, and that is where my diagram starts to break down if you examine it closely enough. At some point, we probably will need to look at several things and decide between collected and distributed. Not that we will want to, but there will be time and technology constraints to consider, and then on top of that many thousands of people that are used to the collected, silo approach to learning will need scaffolding to get out of that mindset.

But the goal is to dismantle, deconstruct, and distribute as much as possible. Unfortunately, I did not capture that well in the diagram.

So let me look at this based on some other questions. Many have asked good questions about how competencies will work across xMOOCs and cMOOCs, as well as how the software for cMOOCs and xMOOCs can be connected in ways that connects designs that don’t fit together. All of these are good questions, and reveal how we will have to look at design in a vastly different way than we do in traditional classes and even many MOOCs.

Most of our design methods follow an ADDIE-ish structure that focuses on defining learner paths, pre-set competencies/objectives, pre-defined outcomes, and collected silo approaches. For this class, my goal is to transition from learner design to learning design. Instead of looking at what each path will look like, what content students will consume, what artifacts they will create, etc, we will need to plop the learner down at the middle of a fluid diagram and then place all of the parts around the learner. Instead of equal and opposite paths with xMOOCs and cMOOCs and instructors and books, we see a learner surrounded with many tools that they can choose as they so desire. Like this:

learning design 1

These parts are all possibilities that the learner can choose from to meet their chosen competencies – but what they use is up to them at any given point (I only put a few examples above – there are many others). Therefore, the design for the class could be more along the lines of giving the learner freedom to choose how they prove they learned what they claim to have learned.

Think of this in the sense of objectives. To have a good objectives, one model is to write a sentence with conditions, behaviors, and criteria. For example:

Using Tableau (CN), the learner will analyze the sample data (B) with at least 90% accuracy (CR).

This is how we typically create objectives in traditional learning. However, the sample data might be meaningless to most learners, and some learners may think 90% is too low while others might be new and think it is too high. To deconstruct this approach to writing objectives, you would start to put in many more blanks:

Using Tableau (CN), the learner will __________________ (B) by ___________________ (CR).

Those blanks would be filled in by students during class. This is the basic idea behind the assignment bank that ds106 uses very effectively. Let the learner set their own objectives and competencies and then let the learner choose how to demonstrate mastery. To do so, the learner might focus heavily on more instructivist resources to meet their objectives, like this:

learning design 2

Or, the learner may focus more on cMOOC resources to meet their objectives, like this:

learning design 3

Notice that not all of the aspect available are used, and even those that are used are not equally utilized. Additionally, these diagrams would not stay static through the whole course – they would morph over time to meet different objectives and competencies as they morph over time. If I had time, this would be better demonstrated by an animation that morphs through several versions as the class progresses. Keep in mind, this is in conjunction with the tubing metaphor of group formation (for those students that choose to have groups) as well as the dual pathways diagram. This is just another abstraction of different aspects.

This would also mean that the various software solutions would only need to be able to connect user accounts across systems and then export artifacts as needed. In other words, we don’t have to worry about how a paper submitted for “Week 1” to EdX can be connected with a blog post for “Week 1” on WordPress or how a discussion response on Facebook can be connected to a Tweet that also responds to the question. We just need to connect the user accounts from EdX and WordPress and Twitter and Tumblr with an account on a central profile (on something like ProSolo). These connected accounts would just need to be able to send out what ever artifact the student wants to use to prove they know the topic. This central profile would not necessarily collect and silo these artifacts, but would link to distributed artifacts. This could be then used for portfolios, badges, certifications, etc. All of this work could be individual or group based (as long as groups come together around shared objectives).

Again, none of this is new stuff – these are all ideas that others have explored. The main goal is to keep this all distributed and open. Anywhere where it doesn’t sound like that is probably just a break down in my explanation. But I need people pointing that out to help clarify and improve the design before we move into production. And I apologize where I had to gloss over some detailed complexities in order to just get to the point. But this is an idea in progress, and I have a bad habit of explaining something and then blowing it up the next day because I found a problem with it :)

Communal Constructivism and Dual Layer MOOCs

Reading through an article by Noel Fitzpatrick & Roisin Donnelly (“Do You See What I Mean: Computer-Mediated Discourse Analysis”), I came across this quote that seemed to speak directly to our current goal of combining xMOOCs and cMOOCs:

“It would seem, at first glance, slightly contradictory to construct communities of practice which are essentially organic structures which should be encouraged to grow, live and die naturally. The challenge is, perhaps therefore, to harness the organic benefits of online communities for sharing and learning within more formalised educational structures.” (p. 6)

This is our challenge – to combine the organic benefits of communities from cMOOCs with the more formalized structure of xMOOCs. It is fairly easy to conceptualize the idea, but the real problem is the community piece. There are really two issues to consider when designing this community: 1) what to do with that content the community creates, and 2) how to organize the communities to facilitate the co-creation of knowledge.

All courses have some type of community that creates some type of content. The real problem with many xMOOCs and most traditional classrooms is that the content created by the community is rarely used by the community (other than “respond to two comments on the discussion board”), and even if used it often vanishes after the class is over. A lot of that has to do with the paradigm: “I’m the instructor, you’re the student, I give the info, you prove you learned it, and then I start over with a new group next semester.” Even in social constructivist designs, the focus may be on what individuals learn and not what they can contribute to the general knowledge in the overall field. Fitzpatrick & Donnelly had another quote that caught my attention:

“With communal constructivism, students and teachers are not simply engaged in developing their own information but actively involved in creating knowledge that will benefit other students and teachers; the focus is on learning with and for others.” (p. 7)

Communal constructivism is not a new term, but one that is often left out of the discussion (except in a few cases). However, the idea is not that foreign since we often see this idea modeled in Reddit. But, of course, Reddit users post links to existing ideas instead of creating their own. What if there was a system like Reddit that connected to blog posts by learners, and then other learners could up vote certain ideas and posts to the level of becoming class assignments or even class content? This would be one way to do something meaningful with the content that the community creates. But really, the broader idea would be to create a system where the focus in on learning with and for others. This is just one idea to accomplish that.

Which, of course, leads to the second issue to consider: how to organize communities. One challenge George Siemens is looking at is how to leverage problem-based learning (PBL) in communities so that these community groupings are not just about discussing content but actually using new knowledge to solve problems. And, if learners have their own data to use, this could also be taken to the level of situated learning.

Of course, most instructors that know what these concepts are also know how to make them happen: create a well-designed PBL lesson and divide students into groups. Except… we’re in a MOOC situation that needs to scale to thousands of learners. This was where my previous idea of cell group models comes in: allow students to self-organize into groups based on whatever groupings they like (geographic, existing relationships, common goal, etc). Or let them sit back and lurk. But each group would identify a volunteer leader to keep things rolling, add new members (if needed), adjust roles when existing members leave, solve problems, multiply the group if it gets too big (and identify another leader), escalate problems they can’t solve, and decide to call it a day when the group organically dies off (even if before or after the end of the course itself).

In order to do this, we would have to stop thinking about group work in the typical ways we usually do: instructor assigns groups, forces all people in groups, and groups never change the whole course. We need to think of these groups in a different paradigm. If you are familiar with river tubing, think of how groups form in that activity. For those who are not familiar, here is a visual:

tubing

As individual tubers float down the river, they drift in and out of groupings as needed. Maybe they stick with people they knew the whole time. Maybe they make new friends as they go and the group grows. Maybe one group wants to go look at something that the others don’t. But the groupings of tubes are constantly changing, growing, and morphing. Some stay solo, some stay in the same group. But as the river (the course) flows, the structure is flexible enough to change over time. (I wish I could find an animated gif of river tubers to better illustrate)

So, layer this flexible grouping system on top of a Reddit system that allows learners to create and up-vote learner content (that can then be used as the problems in PBL), and allow the groups to form around these ideas. It would kind of be like tubers that discuss what drinks to bring, then putting the ones that won the vote into coolers, and those various coolers end up being what draws in the groups of tubers (which would of course change and morph as they float down the river).

Like I have said before, conceptualizing this is easy, but making it happen is a beast to be tamed. Obviously, there are many ways to rig together a system of cloud services and open source installations to make this work. But this is a data analytics course… so we will need to collect data in order to practice what is being preached. Rigging a system together would mean losing a lot of data (and what we get may not interface easily), so at some point we need to find some programming funds to create something.

But this leads back to the xMOOC layer. So far, all of this is mainly focusing on the cMOOC layer in a dual layer course. In  many ways, the xMOOC part is easiest to figure out in and of itself – EdX does a great job of creating a platform for instructivist transfer of knowledge (that many learners want). The goal of a dual layer MOOC is to create multiple pathways through learning, so that students who feel they need more interaction can get more interaction, and those that feel they need more instruction can receive that. Or, those that want to change between the two once or multiple times can. That part is the tricky part. Sure, we can just design lessons that match up and give learners access points to easily change… but again, what about the data? Why did those learners decide to switch at that point? There needs to be a system that collects data on the why and not just the clicks. We need a system that can collect everything from possible frustration to qualitative reasoning… but again, not by going to third party solutions that might lose some data (or not be able to connect the data that is collected because the collection schemas were too different).

Of course, a lot of this mirrors what Dave Cormier has already blazed a trail into with community management. So I can’t claim much of this as my own ideas. Except for the tubing metaphor… because I am Texan and tubing is the state “lazy afternoon activity” of Texas.

Another level of community we need to consider is creating a personality for the course in the way that Jim Groom has mastered. That’s probably going to be a whole other post, but all courses have a personality – the question is are you using yours in a way that engages students or just says “I’m a boring instructor with no personality that just does what millions of other instructors have done before me.”

These are just some ideas for how to solve the challenges before us. I am sure there are other better ones out there. What will we end up doing? I guess you can sign up for the actual MOOC itself to see how we all figure this idea out. If we figure it….

Why Design a xMOOC / cMOOC Hybrid? LTCA Theory

So a lot of interest in the earlier post about creating a dual-layer cMOOC/xMOOC, as well as some of the inevitable backlash. The biggest question rattling around seems to be “why?” Well, my first response is: ask George Siemens – this is all his crazy idea. But I wouldn’t be blogging about the idea and sticking with the team if I didn’t think there was something to the idea. We may run into a huge road block down the road and decide to ditch the idea. But the conceptual part of it is fascinating.

I think some of the initial confusion over the idea stems from the divide between theorists and practitioners. As much as I love theory, many theorists tend to get a little too “either/or” minded for practicality. Its either quantitative or qualitative. Its either behaviorism or constructivism. Its either xMOOC or cMOOCs. And so on. In a practical sense, learning never falls along such clean, neat lines. One moment you need to transfer your expert knowledge to a blank slate, the next you need to let your students struggle and construct meaning from chaos because there are just things you can’t copy and paste into their brains.

In my Ph.D. pursuits, I have been exposed to a new emerging theory called Learning and Teaching as Communicative Actions. This theory is being created by Dr. Scott Warren at the University of North Texas based on the works of Jurgen Habermas. Without communication, learning can not happen. LTCA theory breaks communication down into four forms present in learning:

  • Normative communicative actions are those that communicate knowledge based on past experiences, such as statements in class instructions that lay out expectations for student activities.
  • Strategic communicative actions are the most familiar educational communicative actions – these occur most often through lectures, textbooks, and other methods where specific reified knowledge is transferred to the learner.
  • Constative communicative actions are debates, arguments, and discourses that allow learners to make claims and counterclaims. Constative communication is also where social constructivism connects with LTCA theory, as students come to agreement over constructed knowledge through these communicative actions
  • Dramaturgical communicative actions are those that allow for expression. Learners can reflect or create artifacts that express the knowledge they have gained as well as who that knowledge makes them as a person.

Just looking at all four, I think it becomes pretty obvious that each one requires different paradigms, different design, and different technology among other issues. Yet, we need all four to facilitate effective learning. Lately I have been pondering whether some of our problems in education stem from us trying to cram all of these communicative actions into one software solution, one instructional style, one epistemology, etc. Then, beyond that, we tell all students that they have perform all four at the same time as the other students, “because its not time for discussion yet!” or whatever it may be.

So, the idea of MOOC layers is really looking at a four pronged approach to the idea of teaching and learning as communicative actions using LTCA theory. Yes, we could insert strategic communicative actions into constative communicative actions as the instructor sees fit – but are we really going to do that for all students just because two need it? What if a student that needs strategic actions could just duck out and receive that instruction without disrupting the flow for those that don’t need it?

So, the idea I am digging into is that strategic communicative actions are the domain of the xMOOC. And no matter how much you love or hate xMOOCs, you have to admit that this is what they do best. Dramaturgical communicative actions would be the domain of the cMOOC, especially if we could use things like assignment banks and blogs and basically the entire A Domain of One’s Own set-up. Constative communicative actions would be the domain of the design of the course, using activity design to encourage students to interact and debate. Normative communicative actions would be a mixture of the profile system that pulls students together in groups to create their own norms and the instructors vision for the content norms.

The imperative here is that all of this must scale to massive numbers. This is MOOC design, after all. I know there are ways to do all four prongs in one class without dividing out  into layers. But that will only work if the class is small enough. Whatever criticism you have of the whole idea of “massive”… I agree. Education always works better with smaller numbers. But that is not the reality we are being dealt right now. More and more learners are being crammed in our classes – and they don’t even seem to care how this affects their education. So until the customer (learners) wakes up and starts demanding smaller classes, we have to start figuring out this scale thing. That is reality we live in, and that is the reality we are trying to figure out the best solution for.

Designing a Dual Layer cMOOC/xMOOC

So the task is to design a MOOC that leverages the best of both worlds – xMOOCs and cMOOCs. George Siemens put together a team to look at this possibility for the next MOOC he is designing, and had a meeting called “Design Jam.” Since he works at my University now, I was able to beg/plead/bribe my way on the team. The biggest thing I learned from the Design Jam?

George Siemens and Dave Cormier bicker a lot. And it is very entertaining.

But aside from that, we have a lot of work ahead of us. The main design issue seems to revolve around having multiple paths through the content, mostly focused on creating a connectivist, learner-centered group work approach for those that prefer it, and also an instructor-centered path that guides the learners through the process for those that want that.

Easy, right?

So the basic idea is that learners would enter the course and be presented with the option of going through one of the two routes. Maybe at some point an Artificial Intelligence data-driven program will even be able to recommend the path for them. Learners would enter one of the two paths and follow the paradigm presented. At any time that the learners on the cMOOC track need help (or at some point, when the AI data identifies a need), they can be directed towards the appropriate part of the xMOOC track for help. At any time the learners on the xMOOC track start to get comfortable with the idea of interacting with others (or the AI data identifies this), they can move into the cMOOC track. These movements could be a one time switch at any point, or a constant movement back and forth depending on the learner. Or the leaner could stick on the track they prefer the most. Or do both. Or lurk on one or the other or both. The system would basically look something like this:

dual-mooc

(edit – not sure why I designed the original image from right to left. But click on the image for a larger version)

The idea is pretty straight forward, at least at a conceptual layer. This is an idea that I have been batting around in my head for a while and that many others at the Design Jam identified.

The technology behind it is another question.

The xMOOC path is pretty much in place with EdX. They have a good module-based system for presenting and assessing instructivist knowledge. Add on top of that they have connected to other systems through single sign-on and they are down with APIs… they have a system that is ready to connect with other systems as well as allow learners to move in and out as need with ease.

The cMOOC system that sits alongside that? That is another beast. Technology exists to create a learner-centered system (see A Domain of One’s Own)…. but how does this scale to possibly tens of thousands of learners?

Dave Cormier spoke of a system of community managers that he has found success with and that reminded me of something I read about the largest church in the world several years ago. This church in South Korea has close to a million members, yet connects every one of them to the community through a system of small groups that they call cell groups.

The idea of cell groups is an interesting one because it is based on the idea of organically formed groups that change, grow, die, combine, and otherwise fluctuate as needed. They can form based on location, shared interests, existing relationships, common goals, etc. The groups basically process the teachings of the church together and what they mean for their lives. If people join the church, they can join an existing cell group or form a new one. If existing ones grow too big, they can multiple into two or more. If a group dwindles, it can shut down and the remaining members can join other ones. Each group has about 8-20 members and one or more volunteer leaders that guide the group and run weekly meetings.

Every 8-20 groups are organized into sections, with the volunteer cell group leaders in the section meeting once month to go over issues. Every 8-20 sections are group together into districts, and so on. After a while this may not apply to education in a course that runs for maybe 5-8 weeks. But the idea would be to create a support system for the volunteer cell group leaders that could be based on, say, Teacher Assistants (aka “section” or “district” leaders) instead.

So the idea could be to organically form cell groups in the cMOOC, with each group forming based on location, shared interests, existing relationships, common goals, or whatever they like and they self-selecting a group leader. Roles for group leaders can be laid out a head of time. These could then be further grouped into sections under TAs as needed to deal with bigger issues that may arise. groups can then work together, grow, multiple, dies off, etc as needed for the life of the course, sending issues to the TAs as needed. Volunteer cell group leaders could meet in groups occasionally or as needed for guidance and help.

The problem comes with the software needed to do this. Kin Lane was brought to Design Jam to discuss APIs, and the idea of learner profiles was brought to the table during this discussion. These profiles could be used to help learners identify interests, goals, relationships, etc. Learners could then use their profiles to start forming groups as well as identifying these groups to the profiling system. The data behind the profile system could also identify potential group members. As groups grow, they could multiply (especially as new learners enter from the xMOOC strain). As other groups dwindle from drop-outs, existing members could use the system to identify new groups to join.

This software would also need to identify and map the cell group system in order to group cell groups into sections. It could also identify outliers that haven’t joined a group and see if there is an issue (some may just want to lurk, but others could be confused).

Further design of this system could even create a system for creating interactive spaces that don’t rely on third party products like Google Hangouts or Skype. Not that those are bad, but a lot of important data could be lost in those systems. If something like WebRTC could be integrated into this API driven profile system, learners could form, interact, and leave groups as needed throughout the lifetime of the course and just use the profile system to interact with video, text, etc through their browsers instead of a third party service. Since the cMOOC, xMOOC, social, and pedagogical systems would all be connected, massive amounts of helpful data could be collected throughout the entire class, further refining the system down the road. What leads to new group formation? What leads to groups dissolving? What leads learners to switch from cMOOC to xMOOC or vice versa? On and on.

This is not a fully realized idea or system. But its like we are working on the Sharknado perfect killing machine combination of xMOOCs and cMOOCs. Interesting stuff.