On-line Training Article

Feature \ National Conference — technology-based training
Courses of action
On-line training cannot become truly effective until traditional teaching methods are abandoned, argues Roger Schank. He proposes a ‘goal-based scenario’ approach, to exploit the medium’s capacity to provide learning through doing.

Organisations everywhere are scrambling to get into technology-based training. They see computers, multimedia and the Internet as vital to this enterprise, but have missed the point about what these tools can add to education.

The type of learning that such web-based training initiatives will provide is a cause for concern. There is a danger that it will be the same old stuff as in the past, only this time without the classroom, the trainer or the junkets, which allow time on the golf course with fellow trainees.

Courses have been with us for so long that we simply accept that they have the structure, length and characteristics they should have, and leave it at that. But this is a profoundly short-sighted attitude. At an editorial board meeting of the Encyclopaedia Britannica (of which I was a member), I once asked the board members whether they would be prepared to triple its size, if cost were not an issue — for example, if its contents were delivered on electronic storage media rather than in expensive bound volumes. The subsequent discussion yielded the response that the encyclopaedia was just the right size, thank you. I told them that they would be out of business in a few years. That conversation took place about six years ago. Now it is time for a similar question in a different arena.

Web-based programmes must differ from current, traditionally delivered training courses for three reasons:
• current courses aren’t very effective;
• the length, material covered and general methodology of current courses were derived from practical considerations that are irrelevant in this new medium;
• it’s on a computer; so, we could do some clever stuff.

As we rush to put programmes on the web, it is helpful to remember how well existing courses actually work. How much do you remember about what you learnt on those courses? You memorise what you need, in order to do well, and then forget it. What we do remember is what we have used since we completed the programme. If we don’t use it, we lose it — and most of what we learnt wasn’t all that usable in the first place.

Another reason why we don’t remember is that we often don’t hear the message in the first place. It is actually difficult to listen, when someone talks. To connect what they are saying to what we are thinking requires us to stop listening, and think. In the worst case, we are thinking about something else entirely. Simply put, lecturing is the antithesis of learning. When the content of the lecture is written, the same thing is true — only more so. Web courses don’t need lectures; they need one-to-one instruction.

We remember in only two ways. First, we remember things, which we ourselves have experienced. Second, we learn from the experience of others. In order to profit from another person’s experience, we need to be able to relate it directly to our own. Doing something right once doesn’t fix it in our memories, so we must do it over and over again. Software which allows repetition allows us to become expert at something. The key is doing. Computers are doing devices, and people learn by doing, not by being told.

With all this in mind, I have noticed a number of features which characterise traditional training courses:
• most lecturers are considered good to the extent that they speak well, tell good stories, and use clear examples;
• most lecturers feel the need to define their terms early on; so, a great deal of what happens at the start of a course is the learning of vocabulary;
• doing, to the extent it happens at all, is really only watching the trainer doing. Actual doing by students rarely happens.

The concept of a course is all wrong. A course is of arbitrary duration. Horses run courses of pre-established distances at racetracks. Students are not horses. Students need to accomplish tasks, and, having accomplished them, they should move on to the next task.

I propose the concept of a goal-based scenario (GBS). This is not defined by length of time, number of lectures, or any other passive measure. It is defined by the tasks accomplished. A GBS contains a clear goal; helps students to play a role in realistic situations, during which they may accomplish that goal; provides access to the knowledge required to achieve that goal; and provides instruction from experts at the time it is needed.

GBSs have no fixed duration. Your training is finished when you can perform the skill you are being trained for. This is determined subjectively for a skill such as driving, and objectively when it can be more precisely defined, such as hitting a target consistently with a bow and arrow. Measurement does not drive instruction in a GBS. There’s no arguing about grades, because there are no grades. You simply show that you can do something, and are then certified as able to do it.

GBS designers must consider natural methods of learning, and how performance objectives can be met within those constraints. The key to this is understanding how memory works. Human memory consists of sets of experiences which are linked according to similarity. We remember a new experience to the extent that we can link it with old ones.

Experiences that are totally out of context may be remembered, but they will not be remembered functionally. We will fail to generalise from these, and have trouble recalling them when we need to. The memories that matter are the ones that link to previous memories, from which we can make new generalisations.

Each new experience is processed by using old experiences as a guide. We learn whenever there are meaningful similarities between experiences, and also when there are key differences that previous memories can’t explain. We learn through this explanation process.

We must create our own explanations to understand the world around us. We do so only when we cannot predict what will happen, based on previous experience. In a sense, then, we learn only when we are confused. Without failure there is no need for explanation. Since failure almost never occurs during traditional courses, we can see why they don’t work so well.

It’s possible to create software, which allows students to augment their memories through experience that relates to past events. Multimedia education works, only if it mimics the human learning process. It must induce failure and allow explanation. This requires simulations of situations that are lifelike enough to make the process educationally valid.

An effective simulation must allow for every interesting path that a user may consider taking. If it allows only one or two answers, it undermines the user’s belief that he is in charge of the situation and is really making decisions. Allowing many choices, and simulating realistic responses to each of these, is the hallmark of a good simulation.

Building GBSs is not easy. Finding out what we know about one skill executed in a given situation is more difficult than it looks. We don’t only need to know the right thing to do; we also need to know all the wrong things that have ever been done — and their consequences. Wisdom is understanding what works, what doesn’t, and why.

The content can be obtained by interviewing all the people we can find, who have ever had experiences which relate to what the student is trying to do in the GBS. These must then be linked into the various choice points in the simulation. A good simulation has many more options than any individual will need, because no two students will make the same errors. Each mistake should be connected to an experience from others who have made similar mistakes. This enables students to reflect on what they have done, and create explanations to use as guides in the future.

Can we develop GBSs? Will web-based training allow us to give up on the traditional concept of a course? Will we stop treating trainees as if they were horses with a set course to complete? Will interactivity ever mean more than allowing a user to make random choices, or click buttons to view the next page? Will virtual learning wind up looking like a list of on-line books to read, with on-line videos to watch, or will this really be the chance to set education on the right course?

People Management — 14 October 1999

Site Map

Close this window
________________
ITS Home page