Distance Learning: Is That All There Is, My Friend?

Writers such as Cross (1992) have warned for a generation that the conventional campus/classroom method of education is not a good fit for adult learners. By its essence, the campus-based educational system requires learners to leave the work environment, or delay entering it, often for extended periods. Degree and credit requirements are defined by the administration and faculty, often based as much on tradition and academic politics as on a need-driven plan; these requirements are agreed to by learners when they “vote with their feet” at registration, often with little real negotiation based on personal need. Thus, however well intended, the curriculum is removed from its application; a common result is that even professional training programs often are vulnerable to criticism that their graduates are ill-prepared for the world of work. Corporate universities using classroom training often emulate this model and have similar limitations, even though the classes are shorter and the topics often more job-related.

By contrast, the research on how knowledge is transmitted in the work place and in life shows a far different pattern of learning. For example, Brown’s early observations (Brown 1988; Brown and Palincsar 1989) of knowledge transmission among copier repairmen, since replicated in a wide variety of work environments, demonstrated how adult learners collaboratively create and transmit knowledge, and how poor a fit the conventional classroom model is for these learners. The inescapable conclusion is that if the lecture-discussion method did not exist, application of the principles of situated cognition and cognitive apprenticeship would never lead us to invent it, let alone to emulate it in distance learning. Instead, we should seek alternatives that conform more readily to the patterns of adult learning. From the work on situated cognition, we know that most lifelong learning is intimately tied to the structure of work and life tasks. It is necessarily episodic and contextualized, and often problem-driven, with a rich mixture of procedural and declarative knowledge acquisition and mentoring (cognitive coaching). Peer interaction is often as important as interaction with an instructor. We need distance education models compatible with this pattern of learning.

Part of the answer is to make these learning environments much more problem-centered than is now typically the case in the classroom or the distance-learning course. Learners need to solve problems that are relevant to their goals. These problems become the central and defining characteristic of their learning experiences. The call for more problem solving in curricula is now quite familiar. Note, however, that much problem solving is well structured and algorithmic, although many important problem-solving skills are ill structured.

However, problem solving is not the sole type of learning needed: cognitive psychologists such as Anderson (1995) argue that solving problems requires a rich store of declarative knowledge (facts, concepts, and principles), in the form of appropriately-synthesized mental models. The declarative knowledge upon which the mental models are based can come from many sources: references, direct instructional lessons, peer and instructor tutoring, and the learner’s own prior knowledge. Thus, a well-rounded learning environment should include both types of problem solving and declarative knowledge acquisition.

Distance learning technologies have the capacity to implement learning models that are flexible enough to accommodate all these types of learning outcomes. At the core of these learning experiences should be ill-structured problem-solving activities, perhaps drawn from work in which the learners currently are engaged. However, often the work environment cannot be counted upon to provide appropriate problems. For example, problems needed for full understanding may be too rare, expensive or dangerous, or (if the learner is preparing for a new endeavor or career advancement) the problems may not have been encountered yet, and may not even be predictable. In these situations, case problems and simulations are preferable.

But it is usually insufficient to concentrate on ill-structured problem solving. If we take seriously the implications of Anderson’s ACT model and other research, and if we add what we know about the importance of group processes and motivational factors, then it becomes clear that supports for problem solving should include a variety of knowledge resources sustained by interaction among members of a community of practice. Typical resources to support problem solving can include systems for on-demand retrieval of facts, concepts, principles, and well-structured procedures, examples, and analytical tools appropriate to the task, such as calculators, data collection and modeling tools, and the like. Accompanying these tools, in a common workspace, should be a subsystem for collaborative dialog among participants who are working on the same (or similar) problem(s). The environment should support knowledge construction, so that enduring knowledge artifacts created by the group are added to the knowledge base for the problem, for later use by the learners.

The environment also should be capable of teaching facts, concepts, principles and algorithmic procedures. These declarative (and well-structured procedural) knowledge components can be taught using tutorial methods, including classical computer-based tutorials, as a pre-requisite or co-requisite to the problem solving experience. Tutorial methods need not (and probably should not) emulate the conventional classroom, however. Commercial work with computer-based learning objects by companies such as NETg has demonstrated the feasibility of embedding small tutorial experiences directly in work environments, to be available on demand. In addition, much declarative and well-structured procedural knowledge does not require a full tutorial treatment; it needs only to be retrieved at the right time, in the right way. Technologies that meet this need include Electronic Performance Support Systems (EPSS) and Knowledge Management (KM) systems. These technologies are designed to provide learning on demand, in task-driven environments. They co-exist well with problem-solving activities, whether in the workplace or the learning space.

Advocates of problem-based learning might disapprove of this vision of the distance-learning environment because it validates the appropriateness of formal, direct instruction, albeit in a supporting role. Regardless of the strengths of problem-based learning, there are times when it is most efficient to learn something by being taught it directly. This kind of learning is most needed when the learner is new to a given subject or field, but it also can be required to some degree at nearly any level of study. Note, however, that this is not a call to emulate the conventional lecture/correspondence course structure. In the problem-centered distance learning environment envisioned here, direct instruction is likely to be much more granular and episodic than is the case in typical academic or training courses, whether found in the classroom or in conventional distance learning environments. For this new environment, the convention of the 45 contact-hour semester course will be replaced by 45 one-hour learning experiences - or even by 90 half-hour ones, or 180 quarter-hour ones. These learning experiences will be indexed to the problems, in response to queries such as, “show me how to…”, or, “I want to know why…”, or, “how can I identify a….”

Additional key components are needed in this distance-learning environment. A major challenge in on-demand learning is to keep the learner oriented with respect to the knowledge structure in use as well as the problem being solved. One way to meet this need is to accompany each problem space with a knowledge map. Each node and relationship in the map has a key fact, concept, principle, or algorithmic procedure. Indexed to each node (or cluster of nodes) are appropriate instructional and reference components. These components might be constructed for the purpose, or they might be indexed from general references, or they might be created as artifacts by the collaborative work group of learners. The purpose of this enriched knowledge network is to allow the learner to gain access to the desired knowledge on demand, by “backward chaining” from the problem at hand, or by “forward chaining” from what he or she knows to a particular goal.

While this learning environment is inherently much less linear than a conventional course, it does not follow that there is no sequence. Sequence is driven, not by topics, but by progression of problems, ranging from “novice” to “expert” levels of performance. The problems would be selected (or constructed) according to the demands of the work environment (present or anticipated), but they also should be designed to cause the learner to engage all the “sectors” of the knowledge map. In managed learning environments, the learners can work progressively through a planned sequence of problems. In unmanaged (spontaneous or self-directed) learning environments, the learners can make their own decisions about what problems to work on when, based on advice from the instructor.

The administrative practices of the conventional course also change. Rather than enrolling in a course, learners join a community of practice that is defined by a common goal, such as a particular career advancement, introduction of a new work pattern or technology, and the like. In a work environment, all the people who work on similar problems, even if they have different tasks and levels of skill, would constitute such a community. In academic and lifelong learning environments, the community would be defined by (usually selfselected) common career or life goals.

Note that, unlike a conventional class or course of study, there is no reason for the community ever to disband, as long as its members continue to have a common goal and a common problem space. Thus, a cohort of learners could come together for what begins as a prework structured learning experience, and they could stay together as “graduates” who go on to encounter “real world” problems, free to engage in peer collaboration and to use the knowledge base of the problem space, for as long as they desire. Given the technologies of asynchronous distance learning, there is no requirement for the community to exist anywhere but on-line, though our experience with mixed models of distance education demonstrates the advantages of periodic face-to-face meeting.

The instructor’s role in this environment changes considerably. The instructor’s prime responsibilities include constructing and maintaining the core problems and the supporting knowledge map, and ensuring the integrity of the tutorial and knowledge resources selected as resources. The instructor also leads work group discussion and establishes group norms for inquiry and interaction, by mentoring. The instructor also plays key roles as a counselor and guide. Finally, the instructor has primary responsibility for assessment and certification events (see below).

The roles of individual learners change as well. As the learners approach “expert” level in the progression of problems, they increasingly are able to assume the role of peer tutor. Eventually, this kind of peer tutoring helps make the learning community self-sustaining: as long as the collaborative work group is active, it will be possible for participants to use it as a resource.

Thus, the community of practice continues indefinitely. Each cohort forms a community, works through a progression of problems facilitated by the instructor, and eventually “graduates” from planned activities to “real world” ones. Learning is continuous throughout the life of the community of practice, by means of additions to the knowledge map(s), new tools, and peer tutoring.

To summarize, the key components of the proposed distance learning environment are a progression of problems (real, case, or simulated), on-demand retrieval and tutorial lessons providing supporting declarative and algorithmic procedural knowledge, a knowledge map showing the overall structure of the problem space, and an ongoing community of learners and (peer) instructors. These environments are formed around a common goal, and can persist as long as cohorts of learners are pursuing that goal.

With the key components of the new distancelearning environment identified, some elaboration is needed on the instructional strategies to be used. Much has been made in recent years of the power of collaborative learning and inductive problem solving. It should be clear from the preceding section that this instructional model is at the core of the distancelearning environment envisioned here. However, the previous section proposes that key components also include knowledge retrieval and direct (tutorial) instruction. This is contrary to much popular constructivist rhetoric, which places problem-centered, collaborative learning in opposition to direct instruction. Such opposition is unnecessary; a few theoreticians, such as van Merrienboer, (van Merrienboer 1997) take the position that the two strategies actually are highly complementary, each with its own role to serve, depending on the type of knowledge to be learned. This is the premise of the distance-learning environment envisioned here.

An example may illustrate the complementarity of problem-based and direct instructional models. Suppose we are preparing a distancelearning environment on statistical analysis, for learners who need an introduction to (or review of) quantitative methods. A conventional “on-line textbook” course would probably have explanations of key concepts (such as standard deviation), exercises involving calculation of standard deviation, and perhaps assignments to study a sample research article or two using standard deviation. A problembased approach might involve detailed examination of a research problem involving standard deviation (perhaps constructed by “reverse engineering” a real research report). Learners would be called upon to discuss the key issues in interpretation of data. They would have to decide what the key questions are, and discover that some of the questions involve standard deviation. Then they would have to research (discover) the principles of standard deviation, and the instructor might even require them to derive the formula.

In practice, however, neither of these approaches is likely to be entirely satisfactory. The first approach probably isolates the learner from the application of the concept, while the second may be inefficient, at best. If we combine the two strategies, a third alternative results. In this alternative, learners begin with a research problem (probably a case problem). Working in teams, they analyze the problem and discover the need for a statistical inference using standard deviation (though they would not know the name of the concept at this point). Now, however, the learners are presented with a knowledge map with “standard deviation” drawn as the central concept. Learners can then select mini-lessons to teach key underlying principles (such as, variance or “why does it involve a square, then a square root?”), as well as a range of examples of its application in a variety of contexts. Throughout the study of mini-lessons, learners are in contact with each other and the instructor. They never leave the context of the problem they are trying to solve. They are encouraged to discuss how they can use standard deviation in their work (for example, in market research or in statistical process control), and the instructor encourages the learners to post for discussion examples of its application or issues related to it that they have encountered. The end result is a problem solved, but also a thorough group mastery of the knowledge map, and addition of the competency of using standard deviation to the competencies and norms of the community of practice—all with the efficiency and effectiveness of direct instruction, coupled with the development of a rich mental model suitable for problem solving.

In collaborative problem-centered environments, a frequent concern is assessment. Assessment is important for summative certification of individual competence, of course, but it is also needed to provide formative feedback to individual learners to guide their progress. In the proposed distance-learning environment, most assessment is performance based. This means that learners are held accountable for what they accomplish and how they do so. Conventional knowledge tests may have a place, but only as a formative aid to self-study of declarative knowledge. Much more important are the work products of individuals and groups of learners. Depending on the structure of the group work, it may or may not be possible to isolate the contribution of each individual. If it is not possible, then the summative judgment can be based partly on group performance, and partly on additional, individual problem-solving in the form of a solo project or exam. In many circumstances, though, the ultimate goal may well be professional certification of some sort, and the usual end-of-course summary judgments embodied in grades are irrelevant, and could better be replaced by a formative report such as a skill profile.

Many characteristics of the conventional academic system have evolved as a matter of administrative and economic convenience, rather than optimum effectiveness and efficiency for learning. Furthermore, even conventional online textbook-style distance learning has a structure of development costs, which place a severe strain on many faculty members operating under traditional college policies. If the proposed distance-learning environment is to succeed, therefore, we must consider how it will work administratively and economically.

We must begin with one basic premise: as with any distance learning, traditional administrative and economic assumptions cannot apply. To apply them would virtually preclude attainment of many of the benefits of distance learning. Among the principal benefits are:

• economies of scale

• greater resources for design and development

• improved access to educational services, leading to increased enrollment

• improved integration between learning and work, also leading to increased enrollment

We can examine how each of these benefits can be obtained.

It is neither feasible nor desirable to assume that all of the knowledge components-or even any of them-should be created by individual instructors. Given the state of the art, development of high-quality assets, whether informational or instructional, require resources that make no economic sense for a single instructor or even a single university (academic or corporate). Therefore, it seems most likely that the information and instructional assets in the distance-learning environment will primarily be selected by the instructor (and by learners) and integrated into the knowledge map. The instructor is likely to create only a very small number of informational assets, and even fewer tutorial ones.

There has been much promotion of the “learning object economy” as a way of creating such economies of scale. Examination of the merits of the concept is the topic for a different discussion, but an important point is that development and integration of a coherent, coordinated set of assets requires considerable coordination and planning. Such a structure can be imposed in a number of ways. For example, discipline-wide or profession-wide curriculum standards bodies might do it (often in conjunction with certification tests), a consortium of major distance education providers might do it, an industry standards body might do it, or a large company providing the assets might create a de facto market standard by means of broad acceptance of their products.

Technology advocates, since the days of instructional radio and television, have often attempted to turn the requirement of economies of scale to advantage, by claiming that their technology will make it possible for the greatest minds at the greatest universities to create the greatest courses, which everyone will then use. The flaw in this argument is that it implicitly assumes that there is such a thing as one best course; it ignores the need to create courses to meet the situations and needs of individual learners.

Because the proposed distance-learning environment eliminates the traditional concept of the course, much of the “greatest course” argument for economies of scale loses its strength. It is possible to create learning objects (whether informational or instructional) according to uniform, high standards. But individual instructors, responding to the needs of their learners, will select and assemble these components in a variety of ways for a variety of purposes. Implicit in the modularity of the proposed distance-learning environment is the principle that there is no one “best” course—though there may be a “best repository” for informational and tutorial resources.

Anyone who has developed even the simplest on-line textbook style of distance learning is acutely aware of the level of effort required. Producing high-quality tutorials, simulations, and informational resources is usually beyond the resources of individual faculty or campuses. The traditional solution to this problem is to fund development through grants. However, this solution is far from ideal. Grant funding is almost always a one-time occurrence. There is no provision for the recurring costs of maintenance, upgrade, technical support, faculty training, and dissemination costs (what in the private sector are called marketing and sales). Consequently, the most typical pattern under the present circumstances is for distance learning assets to be adopted only by a small number of faculty, and for the assets to go out of date or become technically unusable in a short period of time. Very few such assets ever make it to “release 2.0.”

In principle, the presence of economies of scale ought to lead to availability of greatly increased resources for design and development of the components of the learning environments envisioned here. But grants usually are not designed to encourage economies of scale. Grants usually value uniqueness and innovation, rather than scalability, broad acceptability and wide use. Granting agencies rarely require needs assessment (market research) beyond the home organizations of the developers. Grants usually require that ownership of assets created be in the public domain, thus virtually precluding ongoing investment from other sources. Thus, grants typically do not create a sustainable model for increased resources for design and development based on economies of scale.

Academic organizations usually are not equipped to create and exploit economies of scale. As argued above, reaching scale requires not only a large investment in design and development to produce high-quality, attractive assets, but also recurring funding to support the ongoing costs of dissemination and development of “release 2.0” and beyond. This kind of sustained and specialized capability can exist within an academic structure only as a separate administrative unit, with its own revenue stream and its own cost accounting; in effect, an independent business unit, capitalized by grants or by the parent institution.

Another alternative is for the organization to exist in the private sector. This strategy has a number of advantages: it allows for venture capital funding rather than grant funding with the concomitant restrictions on scalability, it resolves the copyright and ownership issues in a way that allows continuous reinvestment in the products and their dissemination, and it allows for creation of a sustainable business model with recurring revenue supporting a long-term organization for implementation and maintenance/upgrade. In many situations, it seems likely, therefore, that the most successful strategy for creation of the proposed distance learning assets and organizations is to do so in the private sector. Note that the private sector company could provide only the components of the learning environment, or the whole learning experience, just as some distance learning companies do now.

The proposed distance-learning environment envisions long-term participation and membership, coupled with episodic learning experiences. This pattern of utilization and its associated costs does not fit well with the usual course and semester structure, nor with the usual tuition-for-credit model most academic institutions use to sell their services to their students. Instead, it seems reasonable that the proposed distance-learning environment should be made available on a subscription basis. Subscription pricing could be tied to level of usage or attainment of objectives such as certification. This subscription model would create the recurring revenue stream needed to realize the scale necessary to attract investment of the resources required for high-quality information and instructional assets.

One of the basic arguments in favor of distance learning is improved access by learners who, for a variety of reasons, cannot avail themselves of conventional campus-based learning. However, for these learners, even the typical semester-long course often represents an impractical and unattractive commitment of time. Adult learners often prefer to commit to learning experiences in small increments, based on a continuous balancing of the perceived cost of the learning experience and its perceived benefit. Thus, it is likely that many small learning experiences are likely to be perceived as accessible when a large, semester, course or degree commitment will seem overwhelming and intimidating.

The proposed learning environment, supported by a subscription model, is likely to overcome many of the barriers to access created by the usual course-semester-degree system. It makes it possible for learners to enter the system at very low personal risk as members of the learning community, and it encourages them to greatly expand their involvement in the system as they receive value from it. It seems likely that this model will greatly reduce the high non-completion rate that plagues many conventional distance-learning courses.

While the goal of conventional distance learning is “learning on demand,” the goal of the proposed distance-learning environment described here is “just-in-time learning.” Adult learners especially require that they see a direct link between their work, their goals and perceived learning needs, and their learning experiences. The link should be present logically, but also temporally. By creating an ongoing learning community with many small episodic learning experiences, it is possible to greatly strengthen the link.

By strengthening the link to work, the community of learners envisioned also becomes a community of practice. Successful communities often become self-sustaining, as they develop a life of their own. This, in turn, will strengthen overall utilization of the learning environment, and encourage continued subscription, even by “alumni.”

Despite an evolutionary history of over 100 years, distance learning is at best approaching adolescence: its abundant energy is sometimes unchecked by mature judgment, and through its sometimes-awkward actions we can glimpse real potential. But there is still much to learn about what actually works in distance learning, and all too often we perceive it only within the frame of our conventional assumptions about what education is and how to do it.

The distance learning environment proposed here represents a systemic change in the way we provide learning to adults. It does not attempt to emulate or compete with the campus experience. It should probably best be viewed as a complementary alternative to such experiences, though it may attract some learners away from the campus and from the distance learning courses that emulate campusbased practice.

It seems logical that the private sector (and perhaps the military) will adopt this model first, because so many of the needed components of the proposed learning environment are already in place, though they usually are not yet melded into a single learning environment. Distance learning, knowledge management systems, communities of practice, certification, learning objects, and subscription models for learning services all have a presence in the private sector. What remains is to tie all these components together into the system envisioned here.

In academic institutions, it is not clear that the conventional institutional structure will be able to accommodate the model proposed here. Instead, as has happened with distance learning of other types, it seems likely that the institutions will choose to create new structures that are independent of the conventional ones. Doing so would be the act of an innovative leader. History shows that such visionary leadership is possible, even in the occasional university.

In McLuhan’s rear-view mirror, distance education based on the learning environment proposed here is literally inconceivable. But if we direct our gaze to the road ahead, much more is possible, and even practical.