Information Design

In C. Ghaoui (Ed.) Usability Evaluation of Online Learning Programs. Hershey, PA: Idea Group Publishing. 2003, 299-312.

Learnability

Philip Duchastel

Information Design Atelier

 

Learnability

The usability of online learning programs can be broken down into two distinct issues: the usability of an online university site and the learnability of the course contents.

The first issue concerns the more usual web site usability questions, such as how easy it is for new visitors to orient themselves and get a good overview of what the online university offers and what is involved for themselves in e-learning. Or again, how easy it is for the actual students to interact with the course materials while engaged in learning. These are the traditional issues of usability, important to consider but hardly the most interesting.

It is the second issue, that of learnability, that is the more pressing for designers and educators. The basic question is this: What makes the content of a site [or of some resource] learnable? Take any one of the many thousands of online learning courses currently available on the web and ask yourself: Does this course seem difficult to learn [assuming you have the proper background for it]? What would improve it? What would the ideal online course in this area look like? These questions all underlie the learnability of the course.

What then is learnability? Could we say that it is defined by successful learning? That would mean that students who study the course thoroughly learn its content, as evidenced on a good test for instance. Or could we say that a main criterion is ease of learning? Meaning that students experience good intellectual flow and enjoy the course.

Both of these factors, success in learning and enjoyment of learning, can be considered criteria of learnability. Are there others? That is the issue of learnability and that is what is explored in this chapter.

The skeptic will immediately insist that learning takes place within a learner and that it is that locus that mainly determines learnability – that is, the curiosity, intelligence, motivation and persistence of the learner. These are what make or break learning. The teaching materials can only go so far, the learner has to make a go of it, make it succeed.

While there is some truth to that view, it is certainly not the full picture, nor the most useful picture. Consider traditional usability in web sites or software products. There too, the user plays a role. If he is very dull-witted, or perhaps too pressed for time [showing a lack of interest], or just resistant to learning the basics [you know, jumping in and thrashing around – as often happens], well there is little scope for success no matter how usable the site or program may have been made.

But we don’t give up on usability because of that. We acknowledge the limitations and make assumptions about both the state of the user and the context of usage, then proceed to design to those assumptions. It may not be comprehensive, but it is practical and that is a useful value.

The point is we do not blame the user for incompetence, for ill-will or for the lack of success of our site or program. We maximize usability, realizing well enough that usability is certainly contextual. The same applies, as it should, to learnability: success in learning can be maximized through the product, over and beyond context issues [or in spite of them].

The product view of instruction is an important one, one that is emphasized in the discussion below. An alternate view is a process one: learning is a process, and so is instruction in the sense of manipulating the situation so as to facilitate learning. The process view is not to be denigrated, but a product view can incorporate processes and has definite design advantages. Leanability, it is argued, is best considered in this light.

Learnability of course materials starts then in an attitude, one of ascribing merit to the design of the materials. We may not be able to do much about the state of the learners or about the context of learning [those being educational issues in the institutional sense], but we can maximize learnability via the design of the materials.

Just consider this basic notion: we can all usually identify poor learning materials, even though we will have trouble in the borderland between good and bad, and additional trouble in explicating clearly what the features of excellent learning materials might be. In sum, therefore, learnability is a very real issue.

 

Leading Questions

The challenge before us is to identify those features of excellent learning materials. What makes something learnable? Very learnable, most learnable?

But first, why is it so difficult to pinpoint these features? What are the deep issues underlying learnability? There are three of them I want to raise and discuss in this chapter. They are learning, design, and curriculum. Each is difficult in its own right and learnability involves considering them jointly – you can see the magnitude of the challenge.

The first deep question is what is learning? The field of learning has long been a core issue in psychology and numerous theories of learning have been put forth in answer. The issue is far from settled, as practitioners such as educators well know. There is acknowledgement of different kinds of learning, with different factors at play, but no large agreement on these or on the overall picture.

The second deep question concerns teaching. How do you design for learning? There are general principles that have evolved over time, codified broadly in what is known as the field of instructional design (Reigeluth, 1999a). But here too, there is hardly agreement. All of course will subscribe to general systems principles like those found in software design or in HCI. All subscribe to the value of usability testing, the trying out of the materials designed with sample students in order to verify the strength of the design and capture any ways of improvement. But given divergences in views of learning, it is natural that hard disagreements will occur here too, in how to design for learning.

The third deep question concerns what to teach, the content. At first thought, you might think that this is an outside issue. That first, you decide what to teach, then only after that, how to teach it, how to design it. Or you might think that teachers and curriculum specialists, or professors and institutions, determine the content ‘to be covered’. That learnability applies to any content, whatever it is determined it should be. But that overlooks the crucial notion that the what and the how of learning are inextricably linked. Just as in communication more generally. An instructional designer must fashion the content as much as the process, in the same way an information designer fashions information well beyond the graphic design aspect. Both are information architects, but that is not yet widely recognized, which creates great difficulties for learnability.

We need to consider each of these three issues in turn in addressing learnability. Given the theoretical divergences that enrich the debate, it is impossible to be impartial or to be too firm in one’s perspective. Any representation of learnability will thus be one among many possible. Still, better one than none.

 

Learning - The CIM model

At its most general, learning is the process of internalizing information in memory, making that information available later on when needed. But learning the names of the bones in the body and learning the principles of acoustics are rather different forms of learning. We learn them in different ways. What are the commonalties? What are the differences?

There are three types of learning, conveniently contrasted in what we can call the CIM model. CIM stands for Comprehension, Interest, Memorizing, these being the three factors involved in the learning process.

Comprehension is based on our ability to reason, to fit things together, to see how they all work together. Comprehension is the process of generating internal models of the world in all its workings, large and small. We comprehend when we see how things fit together, how it all makes sense. Learning how the ear processes sound, for instance, involves building a cognitive model of how the ear’s structures handle the acoustic vibrations that fall on them. Understanding is a process of rational model building.

Interest, the second element in CIM, is the attentional factor in learning. If something stands out from its context, it will be more easily remembered, as will things that are extremely vivid or of great personal importance. More usually, we try to learn things that are only of mild interest and then, if attention wanders, learning suffers. Interest has the function of keeping us on task.

The third element, memorizing, handles things that do not fit well together, that have no basis in rationality. The name ‘cochlea’ to represent one of the components of the ear is quite arbitrary to us – there is no reason for it [no reason that we know]. It is [to us] purely arbitrary and no amount of reasoning will assist in ‘understanding’ it. We just have to associate the name and the component.

These then are the elements of learning, each contributing in its own way to capture information and internalize it in memory. Each has certain ways to run smoothly as a process. Comprehension can be enhanced by varying perspective, by seeing something from a different angle. Interest can be engaged strongly through interactivity, a technique much used in many different forms of instruction. And memory of arbitrary elements can be solidified through repeated practice. At a general level, these PIP strategies [Perspective, Interactivity, Practice] are the processes engaged in to provoke learning. We can do it ourselves if we have the initiative, or we can let educators create the learning environment that will do it for us through various forms of instruction.

What of the validity of the CIM model? There is a very general and traditional description of learning as simply the acquisition of knowledge and skills (a unitary definition of learning), but we quickly realize it is too general when we consider processes. Learning the multiplication table and learning to solve mathematical equations involve different, incompatible learning processes.

While there are different types of learning, there is no agreement on exactly what these types are (Duchastel, 1998). The view of comprehension and memorizing as two important types has been long accepted in educational psychology, often called meaningful vs. rote learning, or cognitive vs. associative learning. But what about interest as a type of learning?

We often consider interest as a condition of learning, or as a factor of learning, or even as a secondary result of learning. But one could say the same about meaningfulness (in the case of comprehension) and of explicitness (in the case of memorizing). Describing learning is indeed tricky.

The case for interest as a type of learning is that it corresponds to a psychological process distinct from the other two types. Comprehending and memorizing are two processes, as is also becoming and remaining interested in something. Considering learning to involve three distinct types also directs our thinking in considering how to design for learning, and hence learnability.

 

Design - The MoCaF model

Based on the CIM model, we can see that there will be three types of elements that are needed within an instructional product: models, cases and facts. Combining these [and any product would have all three] leads to the acronym MoCaF for the design model appropriate for the creation of highly learnable instructional products.

Models are the tools of understanding, they are what lead to comprehension. Cases are the illustrative materials that instantiate the models in particular settings. They are the main means of grabbing and holding attention. As for facts, they are just the basics that need to be brutely memorized.

Let us look at each of these three components in turn. I will illustrate them with a discussion of what might be involved in a medical instructional product dealing with hearing and the ear.

1. Models

Models are what drive comprehension. The aim of design in this area is to create models that embody the disparate elements of content while synthesizing them in an artifact [the model] that clearly communicates and is easily learnt. Models show how elements relate to one another; they capture relationships and interactions.

The craft of developing models is one of establishing the underlying structures in a field [content expertise is essential here] and of then representing those structures in synthetic form that facilitate communication and understanding (Wurman, 2000).

Graphic design is a field that has very similar goals when it deals with explanatory designs. A well-known case, not dealing with instruction but with navigation, is the design of the London underground [subway] map, as are subway maps generally speaking. All these are models in that they structure the appropriate information in abstract designs that simplify in order to better communicate and be understood.

In a medical course about the ear, a number of such models are needed to cover the student’s needs. The following models can be seen as appropriate:

These six models are interesting in more ways than one. First, notice the parallels between the first two and the next two – models of normalcy and abnormalcy respectively. Then, consider the last two: models of action to redress a situation [that being the aim of medicine] – the first four models representing what is, rather than what to do. In that category, we also have an internal distinction [1 vs. 2 and 3 vs. 4] between structure and function.

These relationships can all be fitted into a table that models the structure of the knowledge in this field.
 
  Normalcy Abnormalcy Correction
Situation Structure Abnormalities  
Function Ill-functioning  
Action   Diagnosis  Therapy

Now, a further interesting point here is that this knowledge model fits not only the ear, but other body parts as well. It pretty well covers all of individual medicine and can become a framework for medical information and medical education as a whole. Generalizing even more, it can be seen as a structure that can be useful in modeling many fields, perhaps most. In particular professional fields, emphasis may be placed in one model or another, but all models are active in their own way.

2. Cases

Cases are the illustrative material in instructional content. They embody the living problems and the living application of the models. They range from simple examples to complex case studies. Of particular interest are those relatively complex cases that mirror difficult real life settings, such as those used in problem-based medical education or in business education.

Cases are multi-functional in an instructional application. At least three functions can be served:

So for instance, medical students studying the ear would be shown actual cases of structural malformations so that these might be recognized later on when patients are examined. Cases depicting patients arriving at the clinic with various symptoms could be posed as problems for the student to solve, either to practice applying the models or to assess the student’s level of mastery.

Developing cases is generally a labor-intensive process, whose value is also often under-rated. If we consider our knowledge as being dimensionalized in terms of abstraction, cases are the concrete instances of what is encountered in the world whereas models are the abstract, simplified representations of those structures and events. A full knowledge of a field involves comfortably sliding along back and forth on the full range of this dimension.

3. Facts

Facts are those ill-fitting elements of knowledge that are considered important to know and that hang out there on their own, only incidentally attached to some model or other.

In the domain of the ear, for instance, terminology is factual, as are various side-effects of drugs used in therapy, as are also the incidence of various types of ear infection. None of these cohere naturally in a model, but yet are essential to know and must be learned.

Facts are simple to state, for instance in a textbook or in a presentation. But that does not ensure they will be learned. While simple to state, facts are hard to enrobe in a context that will make them easily learnable. As indicated in the section on learning, practice or an eventful context is needed. There are means to accomplish this in instructional terms, such as through games, problems, contests, high-impact media, etc. Often, though, when these are not developed, the student is left pretty much to his or her own devices for rehearsing the facts to be learned. This is not an optimal situation.

Much more can be said about models, cases and facts, and about how to design instructional applications to enhance their learnability. More will be said in the next section in dealing with content, but only so much, as this is a broad topic in itself and needs separate and more elaborate treatment elsewhere.

 

Content

Learning involves interacting with information. And so it is the design of that information that is crucial to learnability. We are dealing here with the content of the instructional product, that content being modeled through design into a certain form that makes it understandable, interesting and memorizable.

It is often thought that the content of an instructional product is given, that it is an entity that pre-exists, and therefore that instructional design is the craft of designing the presentation and surrounding the content with activities that will help the student acquire the content.

This is an unhealthy view of instructional design because that initial assumption about content is wrong. The content is not there simply to be communicated. Each actual representation of the content is an expression of it, a way of putting it. There can be no content without expression.

Every depiction of instructional content, whether in formal fields like the scientific ones or more interpretive ones like the humanities, is a model. Each is a statement of the content at a certain level of complexity. Every model is a simplification, an overlay on reality to help bring out its structure or functioning.

The design of those models is crucial to learnability. It is what will make the content learnable – easy or difficult or impossible to learn. The models are what anchor the entire content of a particular field in an overall coherent structure that the student can assimilate and then use beneficially in building his or her own cognitive structures to represent the field internally.

Models essentially provide structure to hang particulars on and so, the main issue with models will be that of levels. Broad models encompass many more specific models, which contain still more detailed models, which in turn relate the particular facts to one another. A definite hierarchy is involved, from the broad simplified models at the top to the detailed complex information at the bottom.

In the medical information of the ear discussed in the previous section, a broad model of medical knowledge is involved at the top level: it relates normal structure and functioning to abnormal forms of them and then to medical action [diagnosis and treatment]. Then, each of these six components forms its own model, for instance a model of the structure of the ear. And each is broken down into further, more detailed models describing more particular and complex elements. Learning about the ear involves principally learning these models.

As mentioned above, instructional design has not traditionally paid great attention to this structural issue, focusing instead on process aspects of the situation. There are, however, two instructional theories that stand out in their attention to the structural modeling issue: Ausubel’s [1968] advance organizer theory and Reigeluth’s [1999b] elaboration theory. Both call for a progression from broad, organizing models to more specific, elaborate ones as a basis for sequencing the learning experience. Both deal directly with models even if they do not use that descriptor.

The detailed content complement to models, which abstract from reality and simplify, are the cases, which provide the flesh of reality in all its complexity. As for the facts, they just hang loose, without rime or reason, but still have to be learned.

Instructional theory has recently focused on cases in a major way through approaches known as problem-based teaching, simulations, and instructional design based on constructivism. These varied approaches encourage the learners to approach complex issues from their own perspective and figure things out while pulling them together. The main value of cases [problems] is seen in the interest factor, accentuated by the complexity and the challenge of the situation requiring personal investment and deep problem-solving.

While cases are necessarily particularistic, their complexity can vary and this type of content therefore requires design. The case has to be simple enough to be tackled and yet complex enough to challenge, those factors being dependent of course on the assumed sophistication of the learner.

Models, cases, facts are all basically information content of particular sorts, information with which the learner will interact during learning. To a very large extent, then, instructional design is mostly a matter of information design, a notion that needs to become widely recognized. Even the more recent notion of interaction design [often applied to web site design or to exhibition design] is largely a matter of information design involving models [structures], cases [events], and facts [impressions]. In sum, learnability must focus primarily on the content to be included in an instructional product, i.e. on information design.

Now, we might ask whether a product or a process view of design is the better one? Instruction can be viewed both ways and instructional design theory needs to consider both as well (Duchastel, 1998). As mentioned there, learning is an interaction with information products and occurs over time in an organized sequence fashioned as an event. All products unfold events, just as all events involve products.

The product view, however, is better for the designer and for the learnability analyst. The product is the artifact that directs the interaction and that is fully examinable as a designed artifact. It is public, while interactions involve private cognitive processes that are harder to capture and more ephemeral and hence difficult to deal with.

The product embodies the learning event and hence is the locus of learnability. It is what we can most easily point to, comment on, and improve; it is that part of an instructional situation that is touchable. It is basically what instructional designers design.

This is not to say that we focus only on content and not on interaction, on feedback, on level of directivity, learner adaptability, etc. – those being essential elements of consideration in the traditional design of instruction. What the product focus means is that all these traditional criteria of educational soundness must be considered as they are incorporated and embodied in the instructional product. The product shapes all these processes [or not, depending on the case].

But the focus on information design is more crucial in another respect. At its core, information design is an abstraction process, one of determining what to include and what to leave out, in effect, determining at what level and how to present specific content. The subway map mentioned earlier is prototypical of all information design in this respect. It simplifies, it codes [color-coding] and it eliminates the unnecessary detail.

The same abstraction process must be engaged by the instructional designer, who before deciding how to present information, must first decide what to present [and what to leave out], i.e. what level of abstraction is warranted in any given model, case, or list of facts. All the details that are mandated by those who control the curriculum may eventually be presented [in further sub-models, etc.], but how the models are designed is an information design issue, a learnability issue. Learning the subway map of your city might be pretty easy by contrast to learning how to deal with the ear and its problems, but both tasks follow similar routes and are made easier by similar processes of information design. There is full scalability of process.

 

Future

Perhaps the strongest thrust of the argument in this chapter is that to take learnability seriously, one must go beyond superficial design issues and tackle deep concerns in still murky theoretical fields. Not an easy process, to be sure, particularly since it goes up against established traditions – ones that we must re-think if we want to advance the study and practice of learnability.

What are these many traditions? The foremost to overcome is the practice mentioned earlier of assuming that learning success is mostly due to the learner rather than to the learning materials. It is never all one or the other of course, but the balance of responsibility is traditionally skewed. The implication underlying the assumption is that since the learning materials are prepared by someone in the know [the teacher, author, professor, in sum the expert], they must be fine. The confusion of content expertise with expertise in learning and design are very evident, but yet the tradition persists in a widespread manner. Accepting learnability as a factor in learning is a first step in removing the confusion.

A second tradition to tackle is the one prevalent in some software circles, that the interface designer can add bells and whistles and otherwise enhance appeal and ease of use, but heaven forbid any transgressions into the actual design of the software. By extension, the instructional designer can come up with novel twists to interest the learner, but heaven forbid any thoughts of restructuring the content. Information architecture teams comprising expertise in varied areas must design from the ground up, in an open fashion, through deep respect for the value of collaboration. Any nostalgia for the Frank Lloyd Wright stance of illustrious creator must be done away with.

A third tradition to overcome is a paradoxical one: that more is better and yet we offer less than needed. This should be read as follows: that more verbose content [explanations, extrapolations, etc.] is naturally helpful, but that illustrations and cases are too expensive to include and are thus left out. This is the ease-of-preparation tradition that assumes wrongly that learnability can be had at cheap cost [the professor will simply put together the course]. Professors may not have the time to do good case studies, but they are often at ease in being verbose.

Three traditions that come in the way of learnability, not because of ill-intent, but simply because any novel scheme to do things differently is … well, different. That after all is what a tradition is. We are on grounds very familiar to usability experts here.

Perhaps the greatest trend emerging in the future with respect to learnability will be the continuing merging of instructional design into information design. As access to information becomes more ready, we will likely see a reduction of our need to memorize arbitrary information beyond the frequently used or crucial to know kind. Our external memory supports will fill the need for the less needed information.

This merging of the two traditionally distinct design worlds [information and instruction] is particularly informative for the learner-control issue in education. Adult learners like to have more control [or like to think they do] over what they learn, how they learn it and when they learn. They operate more in an access mode than in a traditionally receptive educational mode.

Well-designed information/instruction products will facilitate this approach, being used at times for informational purposes and at other times for instructional purposes. Informal learning [outside of academic structures] and formal instruction both involve learning, both involve interacting with information, both profit from good information design.

The issue of learner control is not an easy one, for it touches on the very nature of instruction as a deliberate attempt to organize learning in a focused way and, at its core, an intentional, albeit benevolent, effort to influence a person’s thinking (Duchastel, 1998). Irrespective of philosophical views, though, it remains that learning will be impacted by this merging of information design and instructional design.

Finally, the issue of generality must be considered. Learnability as discussed here applies to all pedagogical contexts. But does each specific one have perhaps additional learnability constraints or requirements? In particular, does online learning have specific requirements all its own?

My answer is generally negative. Any specific situation will have constraints and needs that are particular to it – a specific anticipated audience, with certain pre-requisites and not others, a controlled time-frame, perhaps pre-dictated media usage or other conditions, and so on. And a designer will naturally adapt the design accordingly.

Concern may be voiced over the lack of attention in this respect to cultural factors and the communicability of the information [its linguistic and more general cultural grounding]. These are important context factors, but we must not overestimate their role in learning. The plasticity of cognition and the general processes of learning are yet more important and give a certain universality to learnability.

While important, these context issues are all nevertheless secondary to the learnability issues discussed above. So while there are additional features to specific instructional situations, it remains the general learnability factors that must be primarily considered in establishing the usability of an instructional product.

This applies to online instruction as well as to more traditional forms. There are technical limitations that flavor the tone of current online instruction, such as information transfer rates, limits to the forms of communication, etc., but all these are but temporary constraints that will be overcome in good time. The design constraints will likewise change with time. What will remain are the general learnability factors.

The design of e-learning materials also might offer more means of controlling interaction than does for instance the design of textbook or other printed materials. This may or may not be an advantage, depending on a whole host of factors, such as maturity of the learners, prior knowledge of the learners and other context factors. But it does raise once again the general issue of content vs. process.

This philosophical issue remains a challenging one, as well as a thorn in any attempt to devise an overarching theory of learning and instruction. One facet of this issue, and I will conclude with it, is why we speak of the learnability of instructional products. After all, we learn also a great deal from interacting with the world at large, not just with artifacts.

The way to come to grips with this issue is to adopt a wide conception of information, as does the field of semiotics. Information goes far beyond the written word, and beyond the world of illustration too. Information is structure that lies within the world around us, both in its structural elements and its processes. Some of these are found elements, others are designed artifacts, ones the design of which we can control. This is where we can affect the learnability of a product or of a structured process.

I hope to have convinced you in this chapter that the usability of learning materials [learnability] is of tremendous value to pursue. Imagine not only how performance can be improved through learnability design, but also how much anguish in learning might be eliminated or at any rate lessened, and what that could do for the practical nature of education. Bringing learnability to this massive enterprise is indeed a worthy cause.

 

References

Ausubel, D. (1968). Educational Psychology: A Cognitive View. New York: Holt, Rinehart and Winston.

Duchastel, P.,  Prolegomena to a theory of instructional design. Online paper and discussion appearing in September 1998 in the ITFORUM and archived at http://itech1.coe.uga.edu/itforum/paper27/paper27.html .

Reigeluth (1999a) (Ed.) Instructional Design Theories and Models: A New Paradigm of Instructional Theory. Volume II. Mahwah, NJ: Erlbaum.

Reigeluth (1999b) The elaboration theory: Guidance for scope and sequence decisions. In C. Reigeluth (Ed.) Instructional Design Theories and Models: A New Paradigm of Instructional Theory. Volume II. Mahwah, NJ: Erlbaum.

Wurman, R. S. (2000) Information Anxiety 2. Indianapolis, IN: Que Publishing.

 

Acknowledgements

I am grateful to comments from colleagues, particularly from Reinhard Oppermann [GMD] and Markus Molz [Universität Koblenz-Landau].