TECHNOLOGY-BASED LEARNING ENVIRONMENTS AS TROJAN HORSES

Elizabeth Murphy

Faculty of Education

emurphy@mun.ca

 

There has been a gradual shift in thinking which is taking much of the focus away from teaching and redirecting it instead towards learning. Along with this shift in thinking, there has emerged an increased interest in the notion of learning environments. For Perkins (1996), an environment has "length and breath, places and parts, non-living and living, simplicity and complexity, constancy and change" (p. v). Wilson (1996) describes a learning environment as "a place where people can draw upon resources to make sense out of things and construct meaningful solutions to problems"(p. 3).

Learning environments are not limited to one type. Besides classroom-based learning environments, we can also talk about constructivist, interactive, multimedia, adaptive, computer-based, asynchronous, constructionist, hypertext or collaborative learning environments. In many respects, there is considerable overlap in the use of terms. Many of these environments are subsets of technology-based learning environments and are therefore variations on a similar concept. Understanding the possibilities and potential that exist with these new environments may represent a first step in evolving a vision for the future of learning in general and for educational reform in particular. As Vosniadou (1992a) posits: " It has become apparent that educational technology may have a better chance to change the school environment if it is based on a new vision of what this learning environment should be".

The purpose of this paper is to explore some alternatives to the traditional classroom-based learning environment through a discussion of three types of technology-based learning environments: Computer-based, Interactive-Multimedia and Adaptive. The aim of the discussion will be to explore the possibilities, potential and, as well, some of the challenges, presented by each type of environment. As such, the discussion will centre around the types of learning afforded by such environments. The technological aspect is de-emphasized in order to first consider the psychological and theoretical underpinnings related to how individuals learn, how they acquire knowledge, and the types of knowledge they acquire in different learning environments. Using this perspective, this paper will explore some of the ways in which technology-based learning environments may act potentially as trojan horses and, as such, drastically alter the educational landscape.

Computer-Based Learning Environments

Several factors have contributed to the developments in computer-based learning environments. Improvements and advances in hardware capabilities have afforded greater computing power. Advances in cognitive and instructional science have moved thinking beyond the limits of behavioural psychology. The new systems of computer-based learning environments are being designed with a view to facilitating complex problem-solving through integrating wholes of knowledge (Dijkstra, Krammer & MerriŽnboer, 1992). Thus, many see in the computer a means to enhance students' cognitive skills and general problem-solving ability. This is in spite of the fact that studies have failed to conclusively confirm the hypothesis that computer-based learning environments facilitate the acquisition and transfer of higher-order thinking and learning skills (Dijkstra, Krammer & MerriŽnboer, 1992). Nonetheless, Salomon (1992) argues that computers make possible student involvement in higher-order thinking skills by performing many of the lower-level cognitive tasks, by providing memory support and by juggling interrelated variables. Through a partnership with the computer, the user may also benefit from the effect of cognitive residue resulting in improvement or mastery of a skill or strategy. Salomon explains:

The intellectual partnership with computer tools creates a zone of proximal development whereby learners are capable of carrying out tasks they could not possibly carry out without the help and support provided by the computer. This partnership can both offer guidance that might be internalized to become self-guidance and stimulate the development of yet underdeveloped skills, resulting in a higher level of skill mastery. (p. 252)

According to Salomon, for computers to have an impact on classroom learning and to effectuate true change, it is the kinds of activities for which they are used and not the computers themselves that must be emphasized. The challenge inherent in this argument, notes Salomon, is that the computer must function like a Trojan horse affording activities that will necessitate profound changes in the learning environment. If computers are to affect learning, they cannot simply serve as an addition to the traditional teacher-dominated classroom in the same way that a television could. Properly introduced, the computer changes the classroom and possibly the school as a whole. For this reason, concludes Salomon: "Computer-based learning environments are not learning environments to which computers have been added (...) Rather, these are relatively new environments in which computer-afforded activities have been fully integrated into other activities, affecting them and being affected by them" (p. 252).

Computer-Supported Intentional Learning Environments (CSILE) (Scardamalia & Bereiter, 1992) provides an example of how computers can significantly alter the educational landscape. CSILE represents a restructuring of the classroom and patterning of schools after scientific research communities in order to provide places for sustained, collaborative inquiry in a "knowledge-building community". The environment is based on the assumption that knowledge is a human construction that takes place as a socio-cultural activity and that it is through apprenticeship with a mature scientist that a young scientist's skills are acquired. Students' work in various academic subjects is entered into a common hypermedia student-generated database accessible to all students so that computer activities are used in all areas of the curriculum.

Jonassen (1992) considers how computer-based learning environments should be designed and varied in order to facilitate advanced knowledge acquisition for expertise in complex and ill-structured domains. Specifically, the design should draw on cognitive flexibility, which is a conceptual model for designing learning environments based on cognitive learning theory. According to this theory, knowledge develops in the three phases of introductory knowledge, advanced knowledge and expertise knowledge. During the first stage, learners have little usable knowledge and instructional systems often do not favour the development of skills beyond reproductive tasks and elemental applications. For this reason, learners often fail to acquire more advanced knowledge. Advanced knowledge is necessary for problem-solving and requires instructional conditions that illustrate the interconnections between knowledge and provide flexible representations of the knowledge domain. In the final stage, expertise is acquired only through domain relevant problem-solving experiences which are actually difficult to capture and make available.

Computer-based instruction has traditionally been characterized by a reductive bias that oversimplifies material and overgeneralizes methods that are more appropriate for introductory knowledge acquisition. As a result, subsequent acquisition of more complex knowledge structures is impeded. Jonassen argues that computer-based learning environments designed with hypertext can support advanced knowledge acquisition in ill-structured knowledge bases. Hypertext is a learner-controlled information base that can adapt its structure to suit the requirements of the learning task, thus allowing learners to acquire more widely applicable and transferable knowledge. Hypertext emphasizes constructivist learning by exposing the learner to multiple perspectives in order to convey the complexity of the content.

Cognitive flexibility theory suggests an approach to the design of computer-based learning environments that is based on understanding how advanced knowledge acquisition and successful learning are achieved. Vosniadou (1992b) also argues for an understanding of the learner and the learning process in order to better design computer-based learning environments. Knowledge acquisition involves restructuring as well as enriching existing conceptual structures. In order for students to be able to restructure these structures, they must first become aware of their entrenched beliefs and be provided with the opportunity to reinterpret them. In the case of science, children often come to the learning task with naive models and understandings of the world based on their everyday experience. These understandings often conflict with accepted scientific explanations. Knowledge acquisition thus requires a radical restructuring of these existing models and understandings. Vosniadou provides an example of such misunderstandings: "...children start with a naive model of the cosmos according to which the earth is flat and stationary and located in the middle of the universe. The sun moves down and hides behind the mountains during the night. The moon has its own light and the stars are small objects found in the sky only at night" (p.150). Vosniadou argues that these misconceptions or entrenched beliefs constrain the types of mental models of the earth that children can form. Factual statements therefore often appear counter-intuitive to students with these entrenched beliefs. For this reason, explains Vosniadou, learning environments must be designed to help students become aware of and question their beliefs, and to help them replace these beliefs with a different explanatory structure. Computer-based learning environments are well-suited to helping students restructure their entrenched beliefs. Computers can model or simulate otherwise highly abstract and unobservable processes thus helping students understand the limitations of their beliefs. Such environments also allow students to create and make external their own representations of systems in order to examine and modify them.

The role of computers in helping students restructure their entrenched beliefs reflects some of the changes which can be brought about at the classroom level. However, more evidence is needed in support of the computer's ability to enhance learning. As well, a better understanding is needed of how the computer might serve as a stimulus or catalyst for broader systemic changes in education in general. Salomon argues in favour of "a systemic, rather than an experimental-analytic paradigm for the study of system-wide changes" (p. 260). Such studies could provide opportunities to qualitatively observe how theories such as cognitive flexibility theory can translate into effective practice in the classroom. Scardamalia and Bereiter (1992) criticize the typical approaches to understanding computer-based learning environments which compare highly experimental innovations with "a stereotype of conventional didactic practice" or studies that typically contrast the "high-tech learner-centred classroom" with the "caricatured no-tech teacher-centred classroom". They argue that more refined contrasts are necessary to better understand the distinctive aspects of the new approaches.

Computer-based learning environments represent complex phenomenon with high potential for improving education and enhancing learning. Our understanding of how such systems can best be designed and implemented remains limited. Effective research in this area will present the challenge of understanding the complex interaction and overlap of variables in the educational setting. Researchers' efforts may be best deployed by first deciding on what types of learning they want such environments to promote. How best to design an environment to support such learning will then need to be considered. It is likely that computers and their software will form only one component of this environment. Roles, curriculum, activities and interactions will constitute some of the other factors that will combine to create the optimal learning conditions in the new environment. As such, understanding computer-based learning environments and their future role in education presents a complex challenge that will require serious and sustained focus and sharing of research efforts.

Interactive Multimedia Learning Environments

Interactive learning is not a new pedagogical approach. Twenty-five hundred years ago, Socrates used it with his students to whom he asked questions in order to promote active thinking on their part. Today, interactive learning, particularly with multimedia, involves considerably more than simply using a Socratic approach. Giardina (1992) provides a description of some of the characteristics of interactive multimedia learning environments (ILMEs). Such environments are not static with fixed roles played by the teacher and learner; rather, the equilibrium is dynamic "where the nature of information and its processing change, depending on the situation, the learning context and the individual needs" (p. v). Wilson (1992) makes reference to the tools and the activities they afford in order to define interactive learning environments as:

...environments that allow for the electronically integrated display and user control of a variety of media formats and information types, including motion video and film, still photographs, text, graphics, animation, sound, numbers and data. The resulting interactive experience for the user is a multidimensional, multisensory interweave of self-directed reading, viewing, listening, and interacting, through activities such as exploring, searching, manipulating, writing, linking, creating, juxtaposing, and editing. (p. 186)

Whereas, at first glance, Wilson's definition may appear to capture the essence of what we might commonly think of as an interactive system, Giardina argues that the notion is actually quite complex and includes diverse elements such as student modelling strategies, multimodal knowledge representation, intelligent advisory strategies as well as diagnostic learning strategies. Another complex issue related to multimediated interactivity is that of control. According to Giardina, there is an intricate relationship between the learner and the environment: "Control and initiative oscillate between the environment and the individual according to the latter's decisions, which in turn are conditioned by the adaptability and flexibility of the environment in relation to individual differences" (p. 48). Depover and Quintin (1992) argue that, in any learning situation, control must be distributed:

Learner control in a learning situation must not be considered as a dichotomous variable expressed in terms of all or nothing but rather as a continuum, from a point of no control, where all decisions would be placed under the responsibility of an external design, to an approach where all decisions would rest in the learner's hands. (p. 235)

Closely related to the concepts of control in learning environments are the notions of adaptability and intelligence. Learning environments endowed with such characteristics might monitor the interactions of the user and react accordingly. Clark and Craig (1992) note that common to most definitions of interactivity is the ability to provide both corrective and informational feedback to the user. A system's ability to provide feedback demands a very sophisticated design since the system must not only be capable of noting the user's actions but of interpreting and then reacting or adapting to them. According to Duchastel (1992), this process of interpretation is difficult for a system to perform. The interpretation is a complex process involving transforming the user's actions into a representation of his knowledge which must be overlaid on the representation of the system's knowledge.

Duchastel explores the possibilities afforded by the merging of intelligent tutoring systems and hypermedia by combining the strengths of both systems into a Hypermedia Intelligent Tutoring System (HMITS). Whereas ITSs structure information and provide intelligent adaptiveness, hypermedia systems do not structure information and contain no pedagogical expertise. On the other hand, hypermedia systems do present considerable potential as learning resources and can allow full learner control relying on the student's own intelligence for learning guidance. HMSs could be designed to provide a level of adaptivity through tailoring of the interface or through selective orientation of the information. A more intelligent HMS would be able to build a student model, didactic knowledge and knowledge of its own display elements. A HMITS would afford learner control while at the same time provide orientation and intelligent adaptability.

Barker (1992) provides an example of a simple adaptive hypermedia interactive environment in the form of an electronic or hypermedia book. Electronic books or e-books can provide the opportunity for the user to experience the material in a variety of formats. Various control mechanisms allow the book to "react" to the interests or needs of the user to explore or view different sections of the book in a non-linear order. Page control mechanisms in the interface allow the user to exit the book, access other books, use related resources. At the same time, control is sufficiently sophisticated to support orientation so that the user knows at all times where he is located in the book. Resources might include "global resources" such as a glossary, notepad or bookmarks. Local page resources could include elements such as pictures, sound effects or video animations or simulations.

Intelligent tutoring systems, hypermedia systems, exploratory, simulated worlds, and e-books are also examples of interactive multimedia environments, thus reinforcing Giardina's affirmation that interactivity is a concept which "readily adapts to all manner of situations". However, as Giardina notes, interactivity is not only defined in relation to the technical considerations of the environment but moreso to the "complex design, the learner's actions and decisions, and adjustments tailored to individual differences"(p. 49). To understand the role that interactivity might play in learning we have first to consider the cognitive dimension and focus on the learning process and not the product (Giardina,1992). Yet, the most important focus in the design or the definition of interactive environments is first and foremost the learner. The concept of interactive learning environments may have become more complex since Socrates' time but the notion of the centrality of the learner has remained an essential and distinguishing feature.

As Giardina aptly argues, interactivity is a complex concept. Control, adaptability and intelligence: these are the three themes noted as central to interactive-multimedia learning environments. But which theme is the most central to the concept? Can a system be interactive if it is not intelligent? Can it be interactive if it does not provide control or does not adapt? Are all systems that provide control interactive? Are all intelligent or adaptive systems interactive? Must all three elements be present for a system to be interactive? In what degrees must they be present? These are but some of the questions that relate to the concept of interactivity.

A central consideration is the distinction between the notions of reactive, interactive and adaptive. To be considered interactive, a system must first and foremost be able to respond to the user. The user's actions trigger a corresponding action on the part of the system. However, what kind of response is required on the part of the system for it to be interactive? If a user presses on the key "w" on the keyboard and a "w" appears on the screen, the system has indeed responded to the user, an action has triggered a reaction, but is this interaction? If, in a hypertext or hypermedia system, a user moves from node to node by selecting links, can the system be characterized as interactive?

Schweir and Misanchuk (1993) refer to levels of interactivity and have constructed a descriptive taxonomy of interaction for multimedia instruction. The three hierarchical levels of interaction are referred to as reactive (in response to a given stimuli), proactive (user generation of unique constructions) and mutual (artificial intelligence). The authors note that a system can incorporate all three levels of interactivity with the mutual approach being of higher quality because of the opportunity provided for "meaningful mental engagement and learner investment". In the mutual level of interactivity, the system adapts to learner progress, advises, assists the user and constructs and refines the environment based on learner input (Schweir & Misanchuk, 1993). Such action suggests both intelligent and adaptive behaviour on the part of the system. Bielawski and Lewand (1991) note that the key factor in intelligent systems is the ability to use knowledge and to associate and infer to perform tasks or solve problems. Adaptability suggests a system which knows something about the user and uses this knowledge to adapt aspects of the system to the user. An example of an interactive system incorporating intelligence and adaptability would be an intelligent tutoring system or expert system. Typically, however, ITSs do not offer a high degree of user control.

The challenge in the design of highly interactive learning environments would be ideally to create systems that can offer intelligence, adaptability and user control. Furthermore, if the interactive system is to respond to a wide range of user's interests and needs, it should be multimodal or capable of a wide variety of media formats. Combining features from different systems to create hybrids such as the Hypermedia Intelligent Tutoring Systems might allow designers greater leeway in combining diverse features to ensure a high level of interactivity as well as learner control. Emerging technologies and those available through the internet such as e-mail, newsgroups, video-conferencing and screen-sharing could also be combined with existing applications and learning environments in order to increase their interactive potential. The concept of interactivity, because it is so complex, affords it the flexibility necessary to merge with a wide variety of systems. At the same time, it can be an over-riding concept - that is- one which drives the design of the entire system. No doubt, as new technologies evolve and emerge, interactivity will continue to be a concept that is privileged by the designers of new systems.

Adaptive Learning Environments

The study of adaptive learning environments links instructional science with computational science. For computational Scientists, artificial intelligence has long been the focus of research efforts. For instructional scientists, the computer is increasingly perceived as a tool for enhancing learning. It is not surprising then that researchers are willing to combine traditionally distinct areas and to engage in interdisciplinary work in order to develop adaptive learning environments. The ultimate goal of such work is to "develop computer systems that provide or support effective learning experiences for a wide range of learners across a broad spectrum of knowledge domains" (Jones, Greer, Mandinach, du Boulay, & Goodyear, 1992, p. 395).

Perhaps the most well-known type of adaptive learning environment is that of the intelligent tutoring system. Such systems are characterized by a knowledge base, a tutoring strategy and, finally, a student model (McCalla, 1992). This student model is what makes an ITS an adaptive system since it is used in order to modify instruction to accommodate the needs of the student being tutored. The system can monitor the student's progress through a particular knowledge base and interpret where the student is and provide feedback on how he or she should proceed. While earlier version of ITSs suffered from a rigidity of prespecified and predictable control paths, more recent ITS research is focussing on creating flexible instructional plans and knowledge bases.

Nonetheless, ITS research still faces considerable challenges in representing large knowledge bases, in varying the student model and in providing a sufficient array of tutoring strategies (McCalla, 1992). An environment integrating so many components into one module represents a highly complex and sophisticated system - one which would be difficult to design effectively. A further weakness of intelligent tutoring systems is that they "fail to consider the context of learning and social interactions fundamental to learning processes" (Jones, Greer, Mandinach, du Boulay, & Goodyear, 1992, p. 384). Other criticisms of ITSs include the fact that they are modelled on an instructionist approach that emphasizes transfer to the student of knowledge that the tutor possesses. du Boulay and Goodyear (1992) question in its entirety the notion of domain knowledge and the ability of ITSs to represent it. They note that a domain such as the history of chemistry is "a human invention". The authors explain: "The danger comes when one starts attributing a 'domain' with an independent existence of its own. One slips from a mode of saying, 'For present purposes, this is how I want to describe the world' to a mode of saying 'This is how the world is'" (p. 321).

The numerous weaknesses and subsequent criticisms of intelligent tutoring systems have led to increased discussion and debate about ways in which they might be better designed. Woolf (1992) provides many suggestions for ways in which ITSs might be improved. She posits that what is needed are better cognitive models or descriptions of learning and teaching which could be encoded into knowledge-based tutors. Work on the modelling of good teachers and subject model experts might compensate for a weaker student model. Understanding of the cognitive processes (such as problem-solving) necessary to accomplish a task and reification of these process in the student and domain model would improve instructional leverage. She argues as well for improved tools that support meta-cognitive activities and that allow the system to infer student intentions from student plans which could be expressed to the tutor.

Other issues to be addressed, according to Woolf, are how often and how much feedback and error correction should be provided to the learner. The communication style of the tutor is also an important area needing to be addressed. Effective systems should be able to maintain interactive discourse with a user and tailor responses to the idiosyncracies of the particular user. At an epistemological level, research efforts will need to consider ways in which to identify the relationships between the system's domain knowledge and what the learner already knows. The tutor should be able to "structure the explanation to follow epistemological 'gradients' along which he or she is likely to comprehend and integrate the new knowledge it contains" (p. 227).

Laurillard (1992) asks how adaptive tutoring systems might be able to diagnose "what a student needs to be taught" (p. 234). Student modelling should enable the system to know the student in the same way it knows the domain. However, Laurillard questions how a model of a student might include aspects such as motivation, perception and interpretation especially when these might vary depending on the task in which the student is engaged. Laurillard also argues that an explicit model of how students learn, while essential to the system, is unknowable. At best, the system can model what students know but not how they came to know it. For this reason, she explains, the student model must remain implicit and "manifested in the way the interaction develops" (p. 246).

Derry (1992) takes a different approach to improving adaptive systems by arguing that we should concentrate research efforts on determining how to design learning environments that favour the development of metacognitive processes. However, Derry is referring not only to the student's intelligence but, as well, to the system's intelligence about itself and its ability to regulate and control its operation. She selects a Vygotskian or cognitive apprenticeship approach to the tutor who serves as a mentor to model appropriate cognitive behaviour for the student. The role of the tutor in modelling support gradually diminishes as the student progresses through the shared activities. The goal of the system is to progressively transfer control of learning from the tutor to the student.

Issues of control have generated considerable discussion in relation to adaptive learning environments (ALEs). Authors Jones, Greer, Mandinach, du Boulay, and Goodyear (1992) point out that, while early systems of the 1970s and 80s gave considerable control to the tutor, today's design trends aim to provide the learner with more tools to control knowledge, reflect on the learning process and inspect what the system knows about the learner. They also argue that shifting classroom roles with teacher as a facilitator must also be reflected in the design of ALEs. The new environments should allow students to construct new knowledge and skills for themselves. One of the most important points raised by these authors is the issue of integrating ALEs into existing school and university environments. They note that the lack of flexibility common to much of school curriculum may make it difficult to introduce new instructional technologies. Restraints on teacher time and training are also factors which may limit the successful integration of the new adaptive learning environments.

Intelligent tutoring systems constituted some of the earliest applications of artificial intelligence in education. However, they have largely been based on traditional methods of learning and teaching and have been implemented primarily in the fields of math and science. They worked well with drill and practice methods with well defined goals for learning such as the development of factual knowledge and procedural skills (McArthur, Lewis & Bishay, 1994). Nonetheless, intelligent tutoring systems are not the only way of conceptualizing and designing adaptive learning environments. Research and developments in the area of online, adaptive hypermedia systems suggest new ways in which technology can be an effective tool for enhancing learning. Intelligent, online help as well as adaptive interfaces may allow for adaptive guidance and presentation based on the user's specific interests and needs. Adaptability may possibly be incorporated into large online hypermedia systems such as the World Wide Web through the use of knowbots or agents. Such systems may allow for a more discovery-type, constructivist approach to learning while at the same time providing some support and guidance to the user.

Progress in the design of ALEs may require investing intelligence in environment tools instead of a tutor. Such an approach might allow for the student to control interaction with the environment through selection of particular tools. However, the degree of student versus tutor control will depend on whether or not the environment is based on instructivist or constructivist principles of learning. For this reason, designers will first need to resolve issues related to philosophy of learning and knowledge before proceeding to more practical concerns related to the development of the environments. What is learning? What is the role of the teacher? What role does the student play in his/her own learning? This are just some of the questions that will need to be addressed prior to designing the environments. As such, the design of adaptive learning environments may need to involve, not only instructional and computational scientists, but cognitive scientists as well.

Conclusion

The metaphor of an environment is intended to evoke a holistic entity comprised of a complex mix of variables that interact, intertwine and interconnect. An environment is an entire amalgam of roles, activities, goals, relationships, interactions, conditions, circumstances and influences that combine to provide the conditions for growth or learning of the individual. Use of technology is but one component of this complex system. Nonetheless, Salomon (1992) aptly argues that technology has an important role to play in this environment and in education in general. Salomon's notion of the computer as the "trojan horse" reminds us that technology is not simply an "add-on", that it has a more complex role to play. That is, if it functions as part of an environment and not simply as another tool to be used when required. That is, if it facilitates a change in other parts of the environment. The trojan horse analogy focuses our attention on technology's hidden capacity to drastically alter the status quo. The analogy suggests an intrusive, subversive approach which is perhaps indicative of the draconian changes that many individuals would hope to see realized in the traditional classroom. In this sense, technology--particularly the newer, emerging technologies--has become like a new-age messiah sent to liberate schools and education from the instructionist, teacher-centered, transmission mode.

The commonality between the perspectives of researchers whose studies were considered in this paper is an encouraging sign. It not only indicates some agreement of the need for educational reform but, as well, agreement on the form that this reform might take. Encouraging, as well, is the obvious preoccupation with starting reform from a discussion of epistemology and of learning theories. Such initial "groundwork" will pave the way for a firm basis for innovation in practice. New educational projects and programs can represent diverse interests and approaches while at the same time have one broad epistemological and philosophical base that unifies and strengthens them.

Whether technology-based learning environments will indeed reveal themselves as the Trojan horse is yet to be seen. It is possible that reform of education will prove to be an illusion or, at the very least, that it will encounter unsurmountable obstacles to its realization. Yet this possibility does not appear to deter researchers, writers and educators in their efforts to design and implement on a small scale their blueprints for new environments for learning. For now, integration of technology into learning makes brighter the proverbial light at the end of the tunnel. One can only hope that such efforts to reform education will not result, instead, in a tunnel at the end of the light.

REFERENCES

Barker, P. (1992). An Object oriented approach to hypermedia authoring. In M. Giardina, Interactive multimedia Learning environments (pp. 132-152). Berlin: Springer-Verlag.

Bielawski, L. & Lewand, R. (1991). Intelligent systems design. New York: John Wiley & Sons.

Clark, R. & Craig, T. (1992). Research and theory on multi-media learning effects. In M. Giardina, Interactive multimedia learning environments (pp. 19-30). Berlin: Springer-Verlag.

Depover, C. & Quintin, J.J. (1992). Learner control versus computer control in a professional training context. In M. Giardina, Interactive multimedia learning environments (pp. 234-247). Berlin: Springer-Verlag.

Derry, S. (1992). Metacognitive models of learning and instructional systems design. In M. Jones, & P. Winne (eds.) Adaptive learning environments: Foundations and frontiers (pp. 257-286). Berlin: Springer-Verlag.

Dijkstra, S., Krammer, H., & MerriŽnboer, J. (Eds.) (1992). Instructional models in computer-based learning environments. Berlin: Springer-Verlag.

du Boulay, B., & Goodyear, P. (1992). Student-system interactions. In M. Jones, & P. Winne (eds.) Adaptive learning environments: Foundations and frontiers, (pp. 317-324). Berlin: Springer-Verlag.

Duchastel, P. (1992). Integrating Hypermedia into intelligent tutoring. In M. Giardina, Interactive multimedia learning environments, (pp. 67-74). Berlin: Springer-Verlag.

Giardina, M. (Ed.) (1992). Interactive Multimedia Learning Environments. Berlin: Springer-Verlag.

Jonassen, D. (1992). Cognitive flexibility theory and its implications for designing CBI. In S. Dijkstra, H. Krammer & J. Merrienoer (eds.), Instructional models in computer-based learning environments, (pp 385-403). Berlin: Springer-Verlag.

Jones, M., Greer, J., Mandinach, E., du Boulay, B. & Goodyear, P. (1992). Synthesizing instructional and computational science. In M. Jones, & P. Winne (eds.) Adaptive learning environments: Foundations and frontiers (pp. 383-401). Berlin: Springer-Verlag.

Laurillard, D. (1992). Phenomemographic research and the design of diagnostic strategies for adaptive tutoring systems. In M. Jones, & P. Winne (eds.), Adaptive learning environments: Foundations and frontiers, (pp. 233-248). Berlin: Springer-Verlag.

McArthur, D., Lewis, M., & Bishay, M., (1994). The roles of artificial intelligence in education: Current progress and future prospects [Avaliable on-line http://www.rand.org/hot/mcarthur/Papers/role.html].

McCalla, G. (1992). The search for adaptability, flexibility, and individualization: Approaches to curriculum in intelligent tutoring systems. In M. Jones, & P. Winne (eds.) Adaptive learning environments: Foundations and frontiers, (pp. 91-122). Berlin: Springer-Verlag.

Perkins, D. (1996). Foreword: Minds in the 'hood. In B. Wilson (Ed.), Constructivist learning environments: Case studies in instructional design, (pp. v-viii). New Jersey: Educational Technology Publications.

Saloman, G. (1992). Effects with and of computers and the study of computer-based learning environments. In E. De Corte, M. Linn, H. Mandl, & L. Verschaffel (Eds.). Computer-based learning environments and problem-solving, (pp. 247-262). Berlin: Springer-Verlag.

Scardamalia, M., & Bereiter, C. (1992). An architecture for collaborative knowledge building. In E. De Corte, M. Linn, H. Mandl, & L. Verschaffel (Eds.). Computer-based learning environments and problem-solving, (pp. 41-66). Berlin: Springer-Verlag.

Schwier, R. & Misanchuk, E. (1993). Interactive multimedia instruction. New Jersey: Educational Technology Publications.

Vosniadou, S. (1992a). From cognitive theory to educational technology. In S. Vosniadou, E. De Corte, & H. Mandl, Technology-based learning environments, (pp. 11-18). Berlin: Springer-Verlag.

Vosniadou, S. (1992b). Fostering conceptual change: The role of computer-based environments. In E. De Corte, M. Linn, H. Mandl, & L. Verschaffel (Eds.). Computer-based learning environments and problem-solving, (pp. 149-162). Berlin: Springer-Verlag.

Wilson, B. (Ed.) (1996). Constructivist learning environments: Case studies in instructional design. New Jersey: Educational Technology Publications.

Wilson, K. (1992). Discussion on two multimedia R & D projects: The Plaenque Project and the Interactive Video Project of the Museum Education Consortium. In M. Giardina (Ed.), Interactive Multimedia Learning Environments, (pp. 186-196). Berlin: Springer-Verlag.

Woolf, B. (1992). Towards a computational model of tutoring. In M. Jones, & P. Winne (eds.) Adaptive learning environments: Foundations and frontiers, (pp. 209-232). Berlin: Springer-Verlag.