Clark Quinn's Learnings about Learning
(The Official Quinnovation blog)

28 January 2015

What I do, don’t do, and why

Clark @ 8:51 am

My background is in learning technology design, leveraging a deep background (read: Ph.D.) in cognition, and long experience with technology.  I have worked as a learning game designer/developer, researcher and academic, project leader on advanced applications, program manager, and more.  More recently, I’ve been working with many different types of organizations including not-for-profits, Fortune 500, small-medium enterprises, government, education, and more with workshops, project deliverables, strategic consulting, writing, and more.

This crosses formal learning, mobile learning, serious games, performance support, content systems, social and informal learning, and more.  I reckon there’s a benefit to 30+ years of being fortunate enough to be at the cutting edge, and I work hard to maintain currency with developments in learning, technology, and organizational needs.I like to think I’m pretty good at it, and I am for hire.  I’ve worked in most of the obvious ways: fixed-fee deliverables when we can define a scope, hourly/daily rates when it’s uncertain, and on a retainer basis to keep my expertise ‘on tap’.

What I have not done, is work on a commission basis. That is, I don’t push someone’s solution on you for a cut of the action. I’ve cut a few such deals in the early days, particularly for long-term clients/partners, but to no avail.  And I’m fine with that. In fact, that’s now my stance.

There are reasons for this both principled, and pragmatic. On principle, I want to remain able to say Solution X is the best, as I truly believe it to be true, and not be swayed that Solution Y would offer me some financial reward.  I believe my independence is in my clients best interests.  This holds true in systems, vendors, individuals, whatever.  I want you to be able to trust what I say, and know that it’s coming from my expertise, not some other influence.  When you get my expert opinion, it is to your needs alone.  And, pragmatically, I’m not a salesperson, it’s not in my nature.

I also don’t design solutions and outsource development. I have trusted partners I can work with, so I don’t need solicitations to show me your skills.  I’m sure your team is awesome too, but I don’t want to take the time to vet your abilities, and I certainly wouldn’t represent them without scrutiny. When I have needs, I’ll reach out.

So I welcome hearing from you when you want some guidance on reviewing your processes, assessing or designing your strategy, ramping up your capabilities, considering markets, looking for collateral, and more. This is as true for vendors as other organizations.  But don’t expect me to learn about your solutions (particularly for free), and flog them to others.   Fair enough?  Am I missing something?

27 January 2015

70:20:10 and the Learning Curve

Clark @ 8:09 am

My colleague Charles Jennings recently posted on the value of autonomous learning (worth reading!), sparked by a diagram provided by another ITA colleague, Jane Hart (that I also thought was insightful). In Charles’ post he also included an IBM diagram that triggered some associations.

So, in IBM’s diagram, they talked about: the access phase where learning is separate, the integration where learning is ‘enabled’ by work, and the on-demand phase where learning is ’embedded’. They talked about ‘point solutions’ (read: courses) for access, then blended models for integration, and dynamic models for on demand. The point was that the closer to the work that learning is, the more value.

However, I was reminded of Fits & Posner’s model of skill acquisition, which has 3 phases of cognitive, associative, and autonomous learning. The first, cognitive, is when you benefit from formal instruction: giving you models and practice opportunities to map actions to an explicit framework. (Note that this assumes a good formal learning design, not rote information and knowledge test!)  Then there’s an associative stage where that explicit framework is supported in being contextualized and compiled away.  Finally, the learner continues to improve through continual practice.

I was initially reminded of Norman & Rumelhart’s accretion, restructuring, and tuning learning mechanisms, but it’s not quite right. Still, you could think of accreting the cognitive and explicitly semantic knowledge, then restructuring that into coarse skills that don’t require as much conscious effort, until it becomes a matter of tuning a finely automated skill.

721LearningCurveThis, to me, maps more closely to 70:20:10, because you can see the formal (10) playing a role to kick off the semantic part of the learning, then coaching and mentoring (the 20) support the integration or association of the skills, and then the 70 (practice, reflection, and personal knowledge mastery including informal social learning) takes over, and I mapped it against a hypothetical improvement curve.

Of course, it’s not quite this clean. While the formal often does kick off the learning, the role of coaching/mentoring and the personal learning are typically intermingled (though the role shifts from mentee to mentor ;). And, of course, the ratios in 70:20:10 are only a framework for rethinking investment, not a prescription about how you apply the numbers.  And I may well have the curve wrong (this is too flat for the normal power law of learning), but I wanted to emphasize that the 10 only has a small role to play in moving performance from zero to some minimal level, that mentoring and coaching really help improve performance, and that ongoing development requires a supportive environment.

I think it’s important to understand how we learn, so we can align our uses of technology to support them in productive ways. As this suggests, if you care about organizational performance, you are going to want to support more than the course, as well as doing the course right.  (Hence the revolution. :)


21 January 2015


Clark @ 8:22 am

In a discussion last week, I suggested that the things I was excited about included wearables. Sure enough, someone asked if I’d written anything about it, and I haven’t, much. So here are some initial thoughts.

I admit I was not a Google Glass ‘Explorer’ (and now the program has ended).  While tempted to experiment, I tend not to spend money until I see how the device is really going to make me more productive.  For instance, when the iPad was first announced, I didn’t want one. Between the time it was announced and it was available, however, I figured out how I’d use it produce, not just consume.   I got one the first day it came out.  By the same rationale, I got a Palm Pilot pretty early on, and it made me much more effective.   I haven’t gotten a wrist health band, on the other hand, though I don’t think they’re bad ideas, just not what I need.

The point being that I want to see a clear value proposition before I spend my hard earned money.  So what am I thinking in regards to wearables? What wearables do I mean?  I am talking wrist devices, specifically.  (I  may eventually warm up to glasses as well, when what they can do is more augmented reality than they do now.)  Why wrist devices?  That’s what I’m wrestling with, trying to conceptualize what is a more intuitive assessment.

Part of it, at least, is that it’s with me all the time, but in an unobtrusive way.  It supports a quick flick of the wrist instead of pulling out a whole phone. So it can do that ‘smallest info’ in an easy way. And, more importantly, I think it can bring things to my attention more subtly than can a phone.  I don’t need a loud ringing!

I admit that I’m keen on a more mixed-initiative relationship than I currently have with my technology.  I use my smartphone to get things I need, and it can alert me to things that I’ve indicated I’m interested in, such as events that I want an audio alert for.  And of course, for incoming calls.  But what about for things that my systems come up with on their own?  This is increasingly possible, and again desirable.  Using context, and if a system had some understanding of my goals, it might be able to be proactive. So imagine you’re out and about, and your watch reminds you that while you were  here you wanted to pick up something nearby, and provide the item and location.  Or to prep for that upcoming meeting and provide some minimal but useful info.   Note that this is not what’s currently on offer, largely.  We already have geofencing to do some, but right now for it to happen you largely have to pull out your phone or have it give a largely intrusive noise to be heard from your pocket or purse.

So two things about this: one why the watch and not the phone, and the other, why not the glasses? The watch form factor is, to me, a more accessible interface to serve as a interactive companion. As I suggested, pulling it out of the pocket, turning it on, going through the security check (even just my fingerprint), adds more of an overhead than I necessarily want.  If I can have something less intrusive, even as part of a system and not fully capable on it’s own, that’s OK.  Why not glasses? I guess it’s just that they seem more unnatural.  I am accustomed to having information on my wrist, and while I wear glasses, I want them to be invisible to me.  I would love to have a heads-up display at times, but all the time would seem to get annoying. I’ll stretch and suggest that the empirical result that most folks have stopped wearing them most of the time bears up my story.

Why not a ring, or a pendant, or?  A ring seems to have too small an interface area.  A pendant isn’t easily observable. On my wrist is easy for a glance (hence, watches).  Why not a whole forearm console?  If I need that much interface, I can always pull out my phone.  Or jump to my tablet. Maybe I will eventually will want an iBracer, but I’m not yet convinced. A forearm holster for my iPhone?  Hmmm…maybe too geeky.

So, reflecting on all this, it appears I’m thinking about tradeoffs of utility versus intrusion.  A wrist devices seems to fit a sweet spot in an ecosystem of tech for the quick glance, the pocket access, and then various tradeoffs of size and weight for a real productivity between tablets and laptops.

Of course, the real issue is whether there’s sufficient information available through the watch that it makes a value proposition. Is there enough that’s easy to get to that doesn’t require a phone?  Check the temperature?  Take a (voice) note?  Get a reminder, take a call, check your location? My instinct is that there is.  There are times I’d be happy to not have to take my phone (to the store, to a party) if I could take calls on my wrist, do minimal note taking and checking, and navigating.  For the business perspective, also have performance support whether push or pull.  I don’t see it for courses, but for just-in-time…  And contextual.

This is all just thinking aloud at this point.  I’m contemplating the iWatch but don’t have enough information as of yet.  And I may not feel the benefits outweigh the costs. We’ll see.

20 January 2015

Getting strategic means getting scientific

Clark @ 8:13 am

I’ve been on a rant about learning design for a few posts, but I ended up talking about how creating a better process is part of getting strategic.  The point was that our learning design has to embody what’s know about how we learn, e.g. a learning engineering.  And it occurs to me that getting our processes structured to align with how we work is part of a bigger picture of how our strategies have to similarly be informed.

So, as part of the L&D Revolution I argue we need to have, I’m suggesting organizations, and consequently L&D, need to be aligned with how we think, work, and learn. So our formal learning initiatives (used only when they are really needed) need to be based upon learning science. And performance support similarly needs to reflect how we process information, and, importantly, things we don’t do well and need support for.  The argument for informal and social learning similarly comes from our natural approaches, and similarly needs to provide facilitation for where things can and do go wrong.

And, recursively, L&D’s processes need to similarly reflect what we do, and don’t, do well.  So, just as we should provide support for performers to execute, communicate, collaborate, and continue to improve (why L&D needs to become P&D), we need to make sure that we practice what we preach.  And a scientific method means we need to measure what we’re doing, not just efficiency, but effectiveness.

It’s time that L&D gets out of the amateur approach, and starts getting professional. Which means understanding the organization’s goals, rejecting requests that are nonsensical, examining what we do, using technology in sophisticated ways (*cough* content engineering *cough*), and more.  We need to know about how we think, work, and learn, and apply it to what we do. We’re about people, after all, so it’s about time we understood the science in our field, and quit thinking that our existing practices (largely from an industrial age) are inherently relevant. We must be scrutable, and that means we must scrutinize.  Time to get to work.


14 January 2015

It’s the process, silly!

Clark @ 8:32 am

So yesterday, I went off on some of the subtleties in elearning that are being missed.  This is tied to last weeks posts about how we’re not treating elearning seriously enough.  And part of it is in the knowledge and skills of the designers, but it’s also in the process. Or, to put it another way, we should be using steps and tools that align with the type of learning we need. And I don’t mean ADDIE, though not inherently.

So what do I mean?  For one, I’m a fan of Michael Allen’s Successive Approximation Model (SAM), which iterates several times (tho’ heuristically, and it could be better tied to a criterion).  Given that people are far less predictable than, say, concrete, fields like interface design have long known that testing and refinement need to be included.  ADDIE isn’t inherently linear, certainly as it has evolved, but in many ways it makes it easy to make it a one-pass process.

Another issue, to me, is to structure the format for your intermediate representations so that make it hard to do aught but come up with useful information.  So, for instance, in recent work I’ve emphasized that a preliminary output is a competency doc that includes (among other things)  the objectives (and measures), models, and common misconceptions.  This has evolved from a similar document I use in (learning) game design.

You then need to capture your initial learning flow. This is what Dick & Carey call your instructional strategy, but to me it’s the overall experience of the learner, including addressing the anxieties learners may feel, raising their interest and motivation, and systematically building their confidence.  The anxieties or emotional barriers to learning may well be worth capturing at the same time as the competencies, it occurs to me (learning out loud ;).

It also helps if your tools don’t interfere with your goals.  It should be easy to create animations that help illustrate models (for the concept) and tell stories (for examples).  These can be any media tools, of course. The most important tools are the ones you use to create meaningful practice. These should allow you to create mini-, linear-, and branching-scenarios (at least).  They should have alternative feedback for every wrong answer. And they should support contextualizing the practice activity. Note that this does not mean tarted up drill and kill with gratuitous ‘themes’ (race cars, game shows).  It means having learners make meaningful decisions and act on them in ways like they’d act in the real world (click on buttons for tech, choose dialog alternatives for interpersonal interactions, drag tools to a workbench or adjust controls for lab stuff, etc).

Putting in place processes that only use formal learning when it makes sense, and then doing it right when it does make sense, is key to putting L&D on a path to relevancy.   Cranking out courses on demand, focusing on measures like cost/butt/seat, adding rote knowledge quizzes to SME knowledge dumps, etc are instead continuing down the garden path to oblivion. Are you ready to get scientific and strategic about your learning design?

13 January 2015

The subtleties

Clark @ 8:06 am

I recently opined that good learning design was complex, really perhaps close to rocket science.  And I suggested that a consequent problem was that the nuances are subtle.  It occurs to me that perhaps discussing some example problems will help make this point more clear.

Without being exhaustive, there are several consistent problems I see in the elearning content I review:

  • The wrong focus. Seriously, the outcomes for the class aren’t meaningful!  They are about information or knowledge, not skill.  Which leads to no meaningful change in behavior, and more importantly, in outcomes. I don’t want to learn about X, I want to learn how to do X!
  • Lack of motivating introductions.  People are expected to give a hoot about this information, but no one helps them understand why it’s important?  Learners should be assisted to viscerally ‘get’ why this is important, and helped to see how it connects to the rest of the world.  Instead we get some boring drone about how this is really important.  Connect it to the world and let me see the context!
  • Information focused or arbitrary content presentations. To get the type of flexible problem-solving organizations need, people need mental models about why and how to do it this way, not just the rote steps.  Yet too often I see arbitrary lists of information accompanied by a rote knowledge test.  As if that’s gonna stick.
  • A lack of examples, or trivial ones.  Examples need to show a context, the barriers, and how the content model provides guidance about how to succeed (and when it won’t).  Instead we get fluffy stories that don’t connect to the model and show the application to the context.  Which means it’s not going to support transfer (and if you don’t know what I’m talking about, you’re not ready to be doing design)!
  • Meaningless and insufficient practice.  Instead of asking learners to make decisions like they will be making in the workplace (and this is my hint for the first thing to focus on fixing), we ask rote knowledge questions. Which isn’t going to make a bit of difference.
  • Nonsensical alternatives to the right answer.  I regularly ask of audiences “how many of you have ever taken a quiz where the alternatives to the right answer are so silly or dumb that you didn’t need to know anything to pass?”  And everyone raises their hand.  What possible benefit does that have?  It insults the learner’s intelligence, it wastes their time, and it has no impact on learning.
  • Undistinguished feedback. Even if you do have an alternative that’s aligned with a misconception, it seems like there’s an industry-wide conspiracy to ensure that there’s only one response for all the wrong answers. If you’ve discriminated meaningful differences to the right answer based upon how they go wrong, you should be addressing them individually.

The list goes on.  Further, any one of these can severely impact the learning outcomes, and I typically see all of these!

These are really  just the flip side of the elements of good design I’ve touted in previous posts (such as this series). I mean, when I look at most elearning content, it’s like the authors have no idea how we really learn, how our brains work.  Would you design a tire for a car without knowing how one works?  Would you design a cover for a computer without knowing what it looks like?  Yet it appears that’s what we’re doing in most elearning. And it’s time to put a stop to it.  As a first step, have a look at the Serious eLearning Manifesto, specifically the 22 design principles.

Let me be clear, this is just the surface.  Again, learning engineering is complex stuff.  We’ve hardly touched on engagement, spacing, and more.   This may seem like a lot, but this is really the boiled-down version!  If it’s too much, you’re in the wrong job.

9 January 2015

Shiny objects and real impact

Clark @ 8:02 am

Yesterday I went off about how learning design should be done right and it’s not easy.  In a conversation two days ago, I was talking to a group that was supporting several initiatives in adaptive learning, and I wondered if this was a good idea.

Adaptive learning is desirable.  If learners come from different initial abilities, learn at different rates, and have different availability, the learning should adapt.  It should skip things you already know, work at your pace, and provide extra practice if the learning experience is extended.  (And, BTW, I’m not talking learning styles).  And this is worthwhile, if the content you are starting with is good.  And even then, is it really necessary. To explain, here’s an analogy:

I have heard it said that the innovations for the latest drugs should be, in many cases, unnecessary. The extra costs (and profits for the drug companies) wouldn’t be necessary. The claim is that the new drugs aren’t any more effective than the existing treatments if they were used properly.  The point being that people don’t take the drugs as prescribed (being irregular,  missing, not continuing past the point they feel better, etc), and if they did the new drugs wouldn’t be as good.  (As a side note, it would appear that focusing on improving patient drug taking protocols would be a sound strategy, such as using a mobile app.)  This isn’t true in all cases, but even in some it makes a point.

The analogy here is that using all the fancy capabilities: tarted up templates for simple questions, 3D virtual worlds, even adaptive learning, might not be needed if we did better learning design!  Now, that’s not to say we couldn’t add value with using the right technology at the right points, but as I’ve quipped in the past: if you get the design right, there are lots of ways to implement it.  And, as a corollary, if you don’t get the design right, it doesn’t matter how you implement it.

We do need to work on improving our learning design, first, rather than worrying about the latest shiny objects. Don’t get me wrong, I love the shiny objects, but that’s with the assumption that we’re getting the basics right.  That was my assumption ’til I hit the real world and found out what’s happening. So let’s please get the basics right, and then worry about leveraging the technology on top of a strong foundation.

8 January 2015

Maybe it is rocket science!

Clark @ 8:10 am

As I’ve been working with the Foundation over the past 6 months I’ve had the occasion to review a wide variety of elearning, more specifically in the vocational and education space, but my experience mirrors that from the corporate space: most of it isn’t very good.  I realize that’s a harsh pronouncement, but I fear that it’s all too true; most of the elearning I see will have very little impact.  And I’m becoming ever more convinced that what I’ve quipped in the past is true:

Quality design is hard to distinguish from well-produced but under-designed content.

And here’s the thing: I’m beginning to think that this is not just a problem with the vendors, tools, etc., but that it’s more fundamental.  Let me elaborate.

There’s a continual problem of bad elearning, and yet I hear people lauding certain examples, awards are granted, tools are touted, and processes promoted.  Yet what I see really isn’t that good. Sure, there are exceptions, but that’s the problem, they’re exceptions!  And while I (and others, including the instigators of the Serious eLearning Manifesto) try to raise the bar, it seems to be an uphill fight.

Good learning design is rigorous. There’re some significant effort just getting the right objectives, e.g. finding the right SME, working with them and not taking what they say verbatim, etc.  Then working to establish the right model and communicating it, making meaningful practice, using media correctly.  At the same time, successfully fending off the forces of fable (learning styles, generations, etc).

So, when it comes to the standard tradeoff  – fast, cheap, or good, pick two – we’re ignoring ‘good’.  And I think a fundamental problem is  that everyone ‘knows’ what learning is, and they’re not being astute consumers.  If it looks good, presents content, has some interaction, and some assessment, it’s learning, right?  NOT!  But stakeholders don’t know, we don’t worry enough about quality in our metrics (quantity per time is not a quality metric), and we don’t invest enough in learning.

I’m reminded of a thesis that says medicos reengineered their status in society consciously.  They went from being thought of ‘quacks’ and ‘sawbones’ to an almost reverential status today by a process of making the process of becoming a doctor quite rigorous.  I’m tempted to suggest that we need to do the same thing.

Good learning design is complex.  People don’t have predictable properties as does concrete.  Understanding the necessary distinctions to do the right things is complex.  Executing the processes to successfully design, refine, and deliver a learning experience that leads to an outcome is a complicated engineering endeavor.  Maybe we do have to treat it like rocket science.

Creating learning should be considered a highly valuable outcome: you are helping people achieve their goals.  But if you really aren’t, you’re perpetrating malpractice!  I’m getting stroppy, I realize, but it’s only because I care and I’m concerned.  We have got to raise our game, and I’m seriously concerned with the perception of our work, our own knowledge, and our associated processes.

If you agree, (and if you don’t, please do let me know in the comments), here’s my very serious question because I’m running out of ideas: how do we get awareness of the nuances of good learning design out there?


Powered by WordPress