Content-type: text/html ~ Stephen's Web ~ In Practice...

Stephen Downes

Knowledge, Learning, Community

Jan 28, 2002

The Panel

SAN DIEGO - I am at the National Learning Infrastructure Initiative conference to take part in a panel discussion titled Learner-Centered by Practice: Applying What We Know About Learning and Cognition in Designing for the Online Environment.

With me were the panel moderator Helen Knibb, David Merrill from Utah State University, and Lynette Gillis, a consultant with Learning Designs Online. Like most quality panels, these members pushed some of my thinking into new directions. And I think as a result we need to restate clearly what we mean by learner centered learning.

So - what do we know about learning? Quite a lot, actually, and we've known it for a long time. Helen Knibb outlined some features:

  • Learning starts from what you already know
  • Learning provides usable knowledge
  • Learning involves learning to learn
  • Learning is community centered
  • Learning is addresses a "discipline base" of knowledge

All this is well and good, but as David Merrill argued, most teaching on campus, online or off, is "terrible." Of 60 online courses reviewed, he argued, only five had any educational value at all. At best, they do nothing more than provide information, but information is not instruction. The net result is what he called "pooled ignorance."

When we more to the concept of learner centered learning, it gets even harder. "Designing open learning is ten times as hard" as designing an online class, he said. Designers must go well beyond traditional "tell and ask" or "simon says" modes of learning.

But that said, good teaching does not change over time. The principles we have understood about learning apply in an online environment. "Technology is just another way to deliver stuff," he said. But the rules don't change: good teaching:

  • facilitates learning
  • apply to any system
  • apply to any architecture
  • are design oriented
Merrill's five principles could form the basis of the Commonplace Book of Learning:
  • Good learning is problem centered
  • It activates previous experience and knowledge
  • It relies more on demonstration than on telling
  • Learners should be required to use their new knowledge of skill to solve problems
  • And it should integrate new knowledge or skills into everyday life

None of the panelists had any disagreement with either Knibb's or Merrill's descriptions of good instruction. But on examination of this concept, especially in a web-based environment, we find ourselves drawn further and further from the traditional model of learning.

Lynette Gillis puts this intuition into concrete form. She described two projects undertaken by her group, one in which call center operators are taught new cell phone features, and another in which staff in a hospital are taught computer systems.

In neither case is the resulting learning structured along the lines of a traditional thirty-nine hour seminar and test. The cell phone application is essentially a virtual cell phone with two modes: show and try. The hospital learning was a combination of computer lab, expert coaching, self-study and documentation.

In preparing for the hospital training, Gillis's group undertook a large scale study looking at preferred training methods based on where participants were in their training. The focus groups suggested that people's preferences might change as they became more familiar with the material, and the study confirmed this suggestion.

When first introduced to a new application, participants preferred a short session in a computer lab. But for follow-up training and support, the vast majority of people opted for experts in the work area to coach them - no more classes. As mastery improved, they began to opt for self-study courses and documentation.

The Issue

As I learn about and experience more and more online learning, I found myself drifting toward a less and less popular position.

Well - that's not completely accurate: the position is one I sketched in my 1998 Future of Online Learning. But while intellectually I felt that it was a good prediction, I am acquiring more of a conviction that it was a good prediction (I won't try to explain the difference; just trust me on this).

The prediction is this: as online learning takes hold, fewer and fewer people will opt for traditional courses and classes, opting instead for less formal learner driven forms of learning.

In my talk I progressed through four phases of increasing evidence for this conviction. Experience - my own experience as an online learner is compelling. I never take classes, and yet have learned most of what I know today - from CGI programming to instructional design to Roman history - in informal, non-structured learning. On reflection, I find myself going through four stages:

  • Theory - I read the background (including even books) or theoretical basis for the discipline
  • Example - I look at examples of what I am trying to study, deconstructing the work, finding out what part does what
  • Practice - I write software, articles (like this one), or create web pages
  • Community - I distribute part of what I create, soliciting feedback, engaging in dialogue, participating in the discipline community

Observations: At dinner yesterday David Merrill suggested that my theory amounts to me wanting everyone to learn the way I do. I think there are worse ways to do it, but not, that's not it.

Having basically lived on the internet since the early 90s, I have had ample opportunity to observe how people who are strongly connected to the internet learn.

these people are not the sort of people who are studied by university instructors evaluating the effectiveness of online learning. For the most part, the university professors never even see this group, much less evaluate them.

For a significant number of people, if they want to learn about something new, they do not sign up for a university course or program, they turn to Google. They find out what information there is about the subject. The internet community expects this of each other (from this expectation comes the expression "RTFM" (Read the, um, manual) before any sort of 'instruction' occurs.

Their second source of information comes from the mailing lists, discussion boards, Usenet posts and other sources of exchange on the net. People new to a field are expected to "lurk" - that is, to listen to the discussion for a bit before jumping in with questions and comments.

Through these discussions learners are exposed to the practice of the discipline. How this works varies with the discipline - programmers and designers are exposed to actual code or designs, writers are exposed to writing and journalists to articles, historians are exposed to the back-and-forth interplay between theorists, and instructional designers to pedagogy.

If the learners need more detailed instruction - and many do - they can opt for it in a variety of ways. In almost every discipline, some sites are dedicated specifically to providing instruction. One example I use a lot in this content is WebReference, a site accessed more than 50,000 times daily by web programmers and designers. Project Cool is another. My own Guide to the Logical fallacies is yet another.

At last, the new learner is expected to put the learning into practice. A new web page design, some sample code, an email question or comment, an article - the essence of the web consists in 'putting it out there.' Seldom does an endeavour fail to elicit comment. And the exchange that occurs between new learner and seasoned professionals integrates the learner into the community.

Research: I could probably refer to Lynette Gillis's work and leave it there. But I am compelled to mention the Open University study that I mentioned last week.

According to the study, the number of young people studying at a distance is increasing rapidly. What's interesting isn't the number of people, it's the reasons they give. They find it cheaper, according to the release, they want to start working right away, and they find the university a bad place to study. "Some of our students have come to the OU having tried studying at other universities, where they have found the lifestyle, including the lack of a strong work ethic that some of them perceive, not for them."

What are we to conclude from this? Minimally, that when given the option, many young people choose not to study in a traditional university environment. That as people become more connected, they choose alternative forms of learning. And that their choices are based in the fact that, on balance, the university isn't really a very good place to learn things, certainly not if you are not prepared to stop everything else in your life for four or more years. And for continued professional development - a rapidly growing area of learning as people become used to living in an era of constant change - a university course or program is simply unreasonable.

Practice: OLDaily is my attempt to instantiate my thinking in practice. Although it is called a newsletter, is designed to be a learning environment. My learners are professional course designers and people interested in online learning. My resources are the materials to which I link every day. My role as an 'instructor' (the word just doesn't fit) consists in my selection of materials and commentary surrounding that selection.

But OLDaily, the newsletter, is just a front. It is the most visible part of what is supposed to be the learning environment as a whole. For as materials are added to OLDaily, they are stored in a knowledge base. the knowledge base is intended to be used as a tool by people working in the field.

My theory of learner centered learning, in short form, amounts to this: learning ought to be created by the learner.

Now let me emphasize that I do not mean 'created' in some sort of constructivist way. What I mean is more like this: where in traditional learning (and traditional online learning) the selection and sequencing of the learning materials is a task performed by the instructor, in learning centered learning the selection and sequencing of the learning materials is a task performed by the student.

The system behind OLDaily, then, is a means of enabling this to happen. When completed (there is still quite a bit of work to do) it should be possible to work the knowledge base in a relatively intuitive way to create a sequence of learning activities directly related to the learner's area of interest.

In the first instance, only one sequence is possible: a chronological sequence. Type something into the search field (or make selections in the advanced search) and a list of resources will be displayed, most recent first (why search engines don't at least attempt to sort by date is beyond me).

The search terms are not matched to the content of the entries in the knowledge base; they are matched to the content of the commentaries. That's important - it allows me to weave threads of thought through the knowledge base by employing a consistent vocabulary. Now what's interesting (in my view) is that I, as the commentator, do not know what threads of thought are being woven into the data. This is something discoverable only by the student.

From today's newsletter, I wrote:

The Parsimony of the Explicit I think there's something important happening in this article but I'm not quite sure I can put a finger on it. Elearningpost summarized it as follows: "David Weinberger: Most Web designers try to control the users? experience. Some try to shape it. And a precious few try to become that experience." That's not bad, but it misses the fact that most of the article is a response against a view proposed by the W3C's Charles Munat to the effect that, "a web site is data, relationships among data, and transformations that may be applied to that data. These are all abstract. For us to interact with a web site, the data/relationships/transformations must become concrete. In an ideal world, the user would have complete control over how this process of um, reification, for want of a better word, occurs." Weinberger's response is, essentially, that if you separate the content from the manner in which the content is presented - that is, if you separate the medium from the message - something important is lost. Now here's the important part: I think that both Weinberger and Munat is correct. The reader must create the relationships and the presentation, and yet, these must also be created by the designer. How how how? Solve this, and you've solved the fundamental problem of learner centered learning.

I think that something like OLDaily is that meeting point. The learning sequence that results is, in a certain way, a dialogue between the instructor and the learner, but one in which the learner doesn't know what the instructor has to say until he asks for it, and the instructor doesn't know what he said until it is asked for.

I've been working with code to take this a step further - the automatic book writer. Now perhaps one day I will write a book, but not soon. One of my major problems with writing a book is that I don't know where to begin and where to end. A book is a linear representation of a three dimensional topic, and thus is never better than a certain facet. And a book is almost always representative of a facet distinct from the learner's interests (which is why the whole concept of sequencing learning objects is odd).

If you would like to try it out, here is my book generator (if you create a book you like, feel free to publish it and give me the royalties - heh).

OLDaily will also include some further features intended to further express this concept. While today there exists only the [Refer] link after each item, I will be adding two more: a [Research] link, which will give the reader an opportunity to create a sequence of resources based on associations with the current resource, and a [Reflect] link, which will give the reader an opportunity to create a running commentary that will be shared with other members of the reading community.

What I am after here is a concept of learning where what is created is an environment, and where learning occurs through working within the environment. It is learner centered not merely in the sense that it is, as David Merrill would say, "open-ended", it is learner centered in the sense that the learning is created by the learner.

Will it work?

It is already working.


I saw this coming but I did it anyways.

In our discussion group here in San Diego - a group consisting mostly of university educators and administrators - I advanced the idea that learners, and not instructors, ought to design their own learning.

Now let me be clear - what I mean by that is learner created learning in the sense I have been describing above. It is not learner centered learning in the sense where we put them into a library and tell them to read. Experts are available. Instruction is available. But the organizing of the learning is undertaken by the learner.

Now in our discussions of learner centered learning the question of learning styles came up - inevitably. Interestingly, to none of us was learner centered learning a matter of designing for different learning styles. Not because we think that people don't have different learning styles (though Merrill hummed and hawed on the point), but because designing for different learning styles doesn't capture the essence of learner centered learning. A good designer can design for different learning styles while at the same time being instructor or institution focused.

But Merrill did raise a division of learning styles advanced by one writer - I forget the name and can't find it on the net (when will people learn to publish their stuff on the net, where people can read it!). Essentially, the author proposed that learning styles can be defined according to the learner's attitude toward learning:

  • Transforming - they engage and transform the learning materials
  • Performing - they do what they need to get an A
  • Conforming - they do what they need to get by
  • Resisting - they do not want to be in this learning situation

There are different ways to approach this thesis. One way is to suggest that you need to design for each of these four groups. Another, volunteered by a member of the audience, is that the objective of instruction ought to be to convert every learner to the transforming mode of learning.

To me, however, the fact that there are people in a class who are not transformers - the fact that there are people in the class who only want a grade, who only want to get by, who don't want to be there at all - is evidence that the learning in question is not learner centered. Or to put the same point another way: learning centered learning means that each and every learner wants to be learning whatever they are learning.

In an important sense, learner centered design is a misnomer. Once we start making decisions for the learner - even if they are in the learner's best interests - we have moved from the realm of learner centered learning to the realm of instructor centered learning.

As I said, I saw this coming. Not exactly Valdy's chorus of boos, but a chorus of comments to the effect that learners are not able to structure their own learning.

I reject that proposition.

Look at it this way: according to David Merrill, the instructors can't teach. And according to Merrill and many others in the audience, the students can't learn on their own. The two sentences can't both be true, because it follows from them that no learning is happening at all, a proposition that is manifestly false.

Now in fact a little of both is happening.

It turns out that university professors, even if they have no schooling in the finer points of pedagogy, are able to convey some knowledge of their field to (at least) those people with a genuine desire to learn about it.

And it turns out that university students, even if faced with "terrible" teaching, are able to organize themselves sufficiently well to be able to learn (and as many commentators point out, a lot of this learning is explicitly informal - study groups, sessions in the pubs, practice in student clubs).

Indeed, it seems to me that most of the evidence and most of the argument against learner centered learning is based on bias in the questions and bias in the practice. Bias in the question, in the sense that self-learners never seem to be included in studies of the effectiveness of learning online. And bias in practice, in the sense that (in a university environment especially) only learning accomplished through formal instructor-centered learning is recognized as legitimate.

I think it would be a useful exercise to develop some tools and some processes to support learner centered learning. To design, for example, a learning management system that sits on the student's desktop and is operated by the student, not the instructor. To compare, in a neutral field, the intellectual achievements of self-managed versus institutionally managed learners. To see what learners can do, if we'd only let them try.

Stephen Downes Stephen Downes, Casselman, Canada

Copyright 2024
Last Updated: Jul 21, 2024 06:08 a.m.

Canadian Flag Creative Commons License.