Start Learning and Stop Being Taught
Seymour Papert was a pioneer in several fronts: AI, math education, robotics (he called it cybernetics). Working at the beginning of the computer era, he saw immediately the immense potential of computer as educational tool and developed the LOGO language to help elementary and middle school students to learn coding, half a century ahead of the current trend. Coding wasn’t a shortcut for a job at Google though; it was, in his view, a way to transform learning at schools. The potential of computer is not to duplicate the pen and paper, lecture and test experience; it is to create a new kind of learning, centered on the learner’s active and syntonic exploration. He vehemently attacked an alternative use of computer at the time, the military style inspired CAI (computer assisted instruction), a computerized drill.
Unfortunately, much of today’s so-called elearning resides mostly in the domain of the latter. Gamification or fancy motion graphics won’t save the day if the learner can see through the façade: it is still a drill, administered by a machine!
We have often talked about the idea of best learning happens when the experience is centered around the learner. But this is not easy. The vast majority of instructional design is still instructor-centered. The very term “instructional design” carries this legacy of military style, strict and assembly line based kind of training.
Maybe back in the 1970s, CAI indeed cannot afford any customization. But today it is not impossible to think of training content tailored to individual needs — it is just more costly! In this article, I will outline some of the ways you can implement learner-centric design.
Choose Your Adventure
If you are delivering training content, it is quite likely that people who need the training play different roles and their needs are different. Following this, the training can be divided into some Essential Tasks and Role Specific Training. For instance, both a regular user and an admin need to learn how to use a software tool for their work. There are common tasks both need to know — we call these essential tasks, or ET. And there are role specific things. How do we divide it up? Instead of going over the essentials first and role specific content with a caveat, we can let the learner identify themselves in the very beginning, and thus assign the main learning path and required “side quests”. These quests will become visible once you reach certain points in your learning, and it only be unlocked if you have identified yourself as a certain role.
The idea of quest is generally speaking appropriate for the learning context. It means three things: 1) learning is organized in a narrative, which is a indeed a privileged form of human perception; 2) learning is modular, divided into small units that each serves unique goals; 3) learning progress is spatial and highly illustrative: the learner starts somewhere and traverses the map along a designated path.
The idea of side quest comes from RPG; it is the idea that apart from the main narrative, there are little activities you can trigger. Every activity has highly measurable metrics: collect 3 of these, or rescue 4 of that. This is excellent for performance-oriented training, where the performance metrics really should be presented as the outcome, instead of certain features of the tool. After you complete those side quests, you will go back to the main path.
Progress can be indicated both by percentage and visually on the map.
Some tasks can be set as optional. These may count as bonus points. This is to be determined by training content.
Who is the Machine? Who’s Human?
Evaluation and assessment are important. This much we know. But when should we do it? Should we offer content upright and wrap up with a quiz, or should we evaluate the learner first, and then offer some customized content?
Ideally, it should be the latter, but in current practice, the majority of cases carry, if the instructional design has been diligent in pushing the SME, some sort of assessment at the end.
This mode of operation comes from classroom teaching: the lecture + test model. It is designed in such a way to produce most efficient learning given the minimal pedagogical resources: one teacher. In the digital age, this is no longer the case, but somehow the bad habits persist.
To treat learners, who are invariably human, as machines who can do nothing but to absorb what is given is a gross mistake. This situation is even more ridiculous given that the only machine is the one giving the lesson.
One thing that helps is to make the learner see that we are treating them as human beings instead of partially complete products going through an assembly line. The unfinished product will be added one part and another regardless of its own wish. This is okay because the products on the assembly line are indeed identical (the exceptions are “defective”). But human beings come through the line with different constituents.
To make a human feel like she is being treated like one, we need to ask her opinion, let her make choices that have visible consequences. I have seen so many courses that start by saying what this course does, and what you, as learner, will get after taking the class. But never for a moment does the course engage with the learner in making inquiries about her special needs. This is wrong. It is a common situation, but it is still completely wrong.
If we are dealing with human beings, this is unforgivable. When you try to talk to someone, the least you can do is to introduce yourself and let your interlocutor do the same. You may not care about what she said and go on lecturing the same content, but this is a nice gesture that makes the other party feel she is being treated as human.
What can we do in the context of elearning? The first such option is to identify the learner’s role. This step accomplishes an important role: to offer a fun, non-obtrusive way to personalize the training content for this specific learner. This is especially useful in cases where learners’ knowledge on the training subject varies a lot.
And the next thing is to choose an avatar. Some might think of it as a gimmick, as it really doesn’t have much impact on the learning content, but this to ignore the psychology of learning: the key is not the content, but how to engage the learner. Choosing an avatar is an important step towards that goal. This is also a special accommodation for diversity.
Your roles may be a given, but your knowledge on the subject is not. We can use a placement test to find out exactly what do you know before hand. However, we do not need to make this explicit, “notice we are going to test your knowledge…” It is psychologically proven that contexts like this will put the learner in an uneasy position and seriously damage their performance. If we are to administer a quiz, make it sound fun and non-consequential. Give the learner more flexibility, instead of the system dictating on everything. The learner may pass a test, but still want to learn more about something. So it is a good idea to let the learner mark a quiz as “I want to know more”. This way the topic is included regardless of the quiz result.
The Garden of Forking Paths
One convenient way that leads to a more learner-centered learning is to create a branching scenario-based learning following the so-called 3C model: challenge, choice, consequence.
Scenario based learning (SBL)itself doesn’t have to be branching. It has its own strength in projecting learners into an imagined yet completely plausible scenario. You can be answering a few questions regarding the scenario, or you can be doing a whole team hands-on project spanning several months.
The branching idea has become a very effective and hence popular form of elearning mechanism. It shifts learner’s attention from passively watching to making active choices; it also shifts their mind from abstract thinking and absorbing knowledge into applying knowledge. Most important of all, it is so EASY to implement in slide-based elearning. Click and jump, done!
However, not all knowledge can be formulated this way. This seriously limits the applicability of SBL to certain areas such as soft skills and safety compliance. It boils down to a sort of psychological test: what would you do under this or that circumstance? One would be hard-pressed, for instance, to come up with a branching SBL for a complex software tool. This is because the expertise of using a software cannot be reduced to some choices — choices do exist at some level, but actually doing them is the real challenge.
Apart from applicability, a branching SBL also has another problem. In a sense, the mechanism is too rigid. A choice you made has an immediate consequence, which is the next question, which is also a challenge. This makes the system’s behavior predictable. In terms of complexity, such a system has little cognitive depth. What does this mean? It means you are asking a simpleton to lecture an adult learner! No wonder they don’t pay attention to you.
Learner Centered, But How?
While everyone would agree that a learner-centered approach is superior, and there are indeed ways to implement this in even the most banal cases of training, in reality, online trainings that deviate from the lecture-quiz model remains scarce. Part of the problem is our understanding of training remains to be updated; and part of the problem is the technological platform. If the elearning authoring tools have made tasks of incorporating active learning easier, I would certainly expect to see more developers using them. On the other hand, if implementing these features takes too much time and money, requires too much coding skills, given the limited resources and tight frame, why expect people to do extra work that is not even appreciated by the boss or client?
