IIMS 96 contents
[ IIMS 96 contents ]

Realism and imagination in educational multimedia simulations

Peter Standen
Edith Cowan University
The power of multimedia to simulate natural and social worlds appears to offer exciting advantages over other teaching methods. However, this is a relatively new application for IMM and little is known about how to develop such simulations. This paper examines some common assumptions about the educational value of realism, focusing on simulations of social contexts that might be useful in teaching business and other applied subjects. It is concluded that there is much to be learned about how to use realism in simulations, and that current methods of developing and studying IMM will not be able to provide answers. Some suggestions for advancing the study of realistic simulations are provided.


The ability of digital multimedia technologies to deliver virtually real worlds - environments that appear and act like the real world - excites many educators and trainers. To some, a future in which such virtual realities can be readily brought into the somewhat unreal context of a classroom seems self evidently to offer better learning opportunities than can he achieved with present teaching methods in which 'theoretical' teaching is occasionally augmented with experience through case studies, videos, industry visits and so on. However little is known about how multimedia simulations aid learning, and developers currently have few reliable theoretical bases from which to work. This paper examines a number of assumptions about realism found amongst developers of IMM simulations, using evidence from literature in a variety of fields. It is argued that before developers succumb to the expensive allure of the virtual world a greater understanding of the role of simulation in teaching is needed.

Learning outcomes in simulations

Little can be said about using IMM or any other teaching method without a clear understanding of what is to be learned. Strangely, discussion of 'learning' in the IMM literature often implies a unitary concept that can usefully summarise the diverse activities of all students in all subject areas in schools and universities. Occasionally distinctions are made between abstract or conceptual "knowledge" and "skills" or procedural knowledge. Common illustrations of principles for IMM design tend to focus on clearly defined conceptual knowledge such as teaching geometry, the laws of physics or on skills such as those used in operation of machinery. In Looking at the uses of simulation in social science subjects, a much broader view of learning outcomes would be appropriate.

This paper addresses the use of simulations in teaching business subjects. In general desirable outcomes might involve understanding theoretical descriptions and prescriptions for practice, but go considerably beyond these in requiring students to learn about the look and feel of the social worlds of practitioners, including cultures, modes of communication, ethical issues and the many ways in which application of theoretical knowledge is compromised by real world constraints of budget and time, as well as by psychological, social and political processes. As theoretical knowledge has a short shelf life and limited generalisability, teachers often also desire outcomes in the learning to learn domain. A brief summary of the outcomes that might be sought in business simulations would therefore involve at least five domains:

A more complete version of this list might include affective outcomes such as positive regard for the subject, and short term motivational goals such as creating a sense of excitement and engagement during learning, although there is not room for detailed discussion of these here.

The major question of this paper is how IMM simulations can promote outcomes of these sorts. While there is not yet any comprehensive discussion of the issues involved in simulating social environments, it is apparent that many developers share common assumptions about the educational role of realism. Here five assumptions will be explored through literature from a variety of areas. These assumptions are: that more realism is better than less; that it is the power to provide rich representations of the world that gives IMM technology an edge; that psychological research shows that realism aids memory; that educational research shows that learning is best situated in realistic contexts and; that realism is a straightforward value free concept.

Assumption 1: More realism is better

Discussions of learning in applied fields often give prominence to the issue of transfer: learning is only considered successful to the extent that it can be transferred to real world problems. It follows that the more teaching involves real world contexts the greater the transfer. In business there is a long tradition of teaching through case studies, rich verbal descriptions of fictional organisations in which principles are illustrated. Students are asked to use imagination to explain or predict the actions of participants in these stories. There is a strong temptation to see IMM as a natural evolution of the case study approach, making it more lifelike through use of pictorial and auditory representations and through interactivity. However a more nuanced version of the view that more realism is better can be gained from two other fields which have confronted the issue.

The view from the training literature

A common theme in the industrial and military training literature since Gagne (1954) is that too much realism is possible. Results of early empirical studies clearly showed situations in which reduced realism would allow learning of essential skills more efficiently (Hays and Singer 1989). Fiske (1983, cited in Sims 1991) points out that training in the real situation is often very inefficient because the real world A further problem is that high degrees of realism are expensive, and a common suggestion is that rather than sock maximum realism training devices should incorporate as little fidelity as will achieve required training outcomes (Hays and Singer 1989). Amongst aircraft simulators there are a number of examples of low or medium fidelity devices that produce equivalent learning to more expensive high fidelity ones (Hays and Singer 1989). Although this literature deals largely with problems of training employees to use equipment such as aircraft, radar systems or industrial machinery it is likely that in social word simulations there are also sometimes advantages to the use of less realistic representations.

A second important point in this literature is that realism is not a uni-dimensional concept. Hays and Singer (1989) divide 'fidelity' into physical and functional dimensions, and discuss a number of related conceptions including equipment, environmental, psychological, behavioural and physical forms of fidelity. Unfortunately, these aspects of realism are poorly defined and even the physical/ functional distinction has not yet been systematically explored in empirical studies.

Although an adequate analysis of the concept of realism in social world simulations is even more remote, designers should consider which aspects of their application may benefit most from realism. It may be, for example, that functional realism is difficult to achieve in the sense of modelling the operation of an organisation or a market, but that students could benefit from high physical realism, conveying the 'look and feel' of an organisation, coupled with a very coarse simulation of its actual operation. In other situations such as teaching financial or economic principles, functional realism might be more valuable than physical.

The view from the simulation and gaming literature

Another view of the role of realism comes from the educational simulation and gaming movement, which has long proposed the benefits of teaching business and other social subjects through such participative and experiential activities. Most of these activities involve functional rather than physical simulations of the world, and it is commonly argued that replication of the world is not the major objective of such games. Rather, students are developing skills in thinking about messy social environments - as Goodman (1995) puts it, they get 'practice in theory'. In addition they often get to work on related 'generic ' skills involved in problem solving, negotiation and teamwork. It is assumed that these skills transfer to real domains, along with some basic experiential understanding of the nature of political, economic or psychological phenomena. Such outcomes are not frequently addressed in IMM simulations. The current focus on these abilities in business education and the promotion of lifelong learning models of tertiary education suggest they should have greater prominence. Rather than looking just to illustrate textbook theories, IMM developers may find value in asking students to develop theories about simulated worlds, or about the relationship of the simulated and real worlds, through reflection and discussion. Opportunities for having groups of students work together - or even take adversarial roles - should be explored.

Assumption 2: Technology can and should take students into almost real worlds

It is obvious but often overlooked in discussions of realism that, with the exception of some very expensive simulators for aircraft, space shuttles and the like, a computer simulates only some aspects of the real world, typically visual and/or functional components. Computer screens have low resolution and small visual angles, depict only two dimensions, and are framed by borders and situated in environments such as classrooms, making visual fidelity a relative concept. With video clips the camera angles and grammatical conventions limit ability to construct mental impressions of environments. Keyboards and mice limit kinaesthetic interactivity.

Are such limitations important? Should we view virtual reality technology, which promises to remove some of these barriers, as heralding a new paradigm for learning in which students can gain even more real experiences in the classroom or home study? There is little research which aids understanding of the limitations of current computer representations.

Possibly, however, concern over these issues is misplaced. Observation suggests that motivated students can ignore them and create quite vivid mental 'worlds' from limited information. Laurel (1993) provides a number of illustrations of such processes. For example in her study of the use of 'Guides' to help navigate databases, on screen pictures and brief outlines of the Guides interests provoked users to create rich mental impressions of the sorts of people these fictional characters might be. The ability of text (as in books) and severely degraded visual images (as in early computer games) to evoke satisfying representations and to assist learning is also obvious. Indeed it might be that requiring students to actively imagine rather than passively view worlds has educational value, even if at the level of improving motivation through more direct engagement of the student.

As Laurel puts it, the problem for IMM simulations may be not so much how to capture the detail of the world as how to engage learners' own creative powers:

Given a multisensory environment that is good enough, people engage in projective construction that is wildly elaborate and creative. And so this turns the problems on its head; rather than figuring out how to provide structure with pleasing emotional textures, the problem becomes one of creating an environment that evokes robust projective construction. (p208)
For IMM developers such ideas caution against automatic investment in technology for realism. Rather we need better understandings of these constructive powers of imagination, and under what circumstances displaying reality is superior to asking students to create it. These understandings will differ for different learning outcomes and subject areas. For example in business some students do not have work experience and cannot reasonably imagine office environments, while those with long but not broad work experience may feel challenged to work in environments that differ from their own.

Assumption 3: Psychological theory shows that realism helps memory

Physical realism involves heavy use of non-symbolic visual and possibly auditory representations. A number of theories developed in psychology are widely held to suggest that more realistic (non-symbolic) representations aid in retaining learning. For example, it is said that pictures or graphical illustrations may create better memory than verbal descriptions (Reiber 1990; Schlecter 1993), that audiovisual materials are easier to code than textual ones (Salomon 1978), and that information presented in both verbal (ie symbolic) and pictorial forms is more resistant to forgetting than purely verbal information (Paivio 1986). The value of mixed mode presentations (eg photographs and audio) and more environmentally realistic experiences have also been espoused by audiovisual researchers since Dale (1954). The issues here are complex and can only be touched on briefly. However some concerns about the value of these theories to developers are worth raising.

First there is not a consensus on how abstract learning and memory are. Many psychologists and workers in fields that overlap IMM such as AI and ITS hold that knowledge is semantic or propositional and ' is stored and retrieved from memory. If learning is relatively abstract, then attention to realism in IMM may be a wasted investment or may even detract from learning by interfering with the abstraction process.

Others propose that knowledge at least in some areas is not separated from its sensory (eg visual) basis, though it may be also recoded into non-sensed formats. For example Paivio's (1986) dual coding theory is widely cited in the IMM literature as indicating that both visual and verbal (symbolic) aspects of information are retained. There is also evidence that physical characteristics of voice are encoded along with linguistic content (Geiselman and Crawley 1983).

A third position is that memory does not involve a separate store of abstract symbolic information, but rather consists only of fairly literal images of episodes linked in ways which allow memory traces to act as though they represent abstract symbolic information (eg Jacoby and Brooks 1984). Parallel distributed processing models (eg McClelland 1994), sometimes described as neural networks, show how the architecture of a storage system can accommodate very detailed 'raw' memory traces and yet allow emergence of more abstract rule governed representations.

Recently a very different position on the issue of memory has been popularised by the situated cognition approach to learning. Drawing on the ecological psychology of J. J. Gibson (1986), writers such as Young (1993) proposes that meaning is created 'on the fly' through the interaction of a perceiver's "abilities" and visual clues or "affordances" in the environment, virtually eliminating the need for theories of memory and mechanisms for encoding and retrieval. Educational prescriptions arising from this view stress the need to situate learning in realistic (including physically realistic) contexts, a view discussed further below.

Thus a range of positions on the role of physical (sensory) detail can be found. Those proposing to base IMM applications on any one of these theories should consider the evidence supporting it, much of which involves very different types of learning to that intended in IMM. Typically in the first three frameworks these are laboratory experiments with highly impoverished stimuli such as single words, isolated sentences or line drawings, presented in large quantities in artificial laboratory settings to students who have very different motivations for learning to students in a typical class. There is very little if any demonstration of the applicability of these theories to learning about complex, fuzzy, non-symbolic problems such as those in social contexts.

For IMM developers perhaps the most interesting product of the debates over learning, memory and representation is increasing evidence that learning is not wholly abstract, as indicated in the last three positions above. While there are inconsistent findings and methodological inadequacies, evidence from a wide variety of sources, including studies of educational video and television, seems to point to improved retention for materials that are realistic in the sense of being highly visual or multisensorially lifelike (eg Hapeshi and Jones 1992).

Most proponents of the use of such materials are referring to their use in communicating ideas or contexts. Another use is to provoke use of internally generated visual imagery in problem solving. For example Reiber (1995) cites a number of examples of the use of visual thinking in science history and suggests that "visualisation appears to be the source of some of our most versatile and robust cognitive tools" (p52), although he provides no evidence on how many great discoveries have been made through verbal or semantic thinking. Despite this, the suggestion that instructional designers should teach visualisation as a cognitive strategy connects with the idea that IMM simulations should allow rich visual environments to replace the need for learners to use their own constructive abilities.

Assumption 4: Learning theory shows that learning is best situated in realistic contexts

Situated learning theory (Brown, Collins and Duguid 1989; Cognition and Technology Group at Vanderbilt 1992) has widely been used to inform IMM development, and suggests a strong role for realism. Its extreme form suggests that all learning should take place in contexts that are physically as well as socially and culturally realistic. Learners are assumed to perceive invariances in these contexts and generate solutions; the teacher's task is to provide structured experiences that help perceive the correct information.

Situated learning appeals to IMM developers as it resonates with a widely held feeling that much tertiary teaching is too abstract and students do not learn how to apply knowledge or develop new knowledge in job situations. This is a common theme in the business education literature and accounts for the popularity of the case study approach to teaching noted above. However, situated learning rests heavily on the psychological theories of situated cognition and ecological perception (Gibson 1988, see Young 1993) noted above. Although there is not room here to discuss these motivations, developers should ask whether this theory applies directly to tertiary teaching in their subject area. A number of considerations suggest that the application of this theory is not yet demonstrated.

Situated learning was derived from studies of learning in non-academic environments, and its principles are most often demonstrated by reference to everyday learning, such as Brown et al's (1989) famous problem of cutting the cottage cheese. Demonstrations of its efficacy appear largely limited to learning through experience. For example in the widely cited Jasper series of problems (Cognition and Technology Group at Vanderbilt 1992), students learn basic properties of the physical world through a series of problems that involve direct experience.

The application of situated learning to tertiary education has been queried by Laurillard (1993), who contends that much of the knowledge of the world taught in universities is "known more through exposition, argument, interpretation; it is known through reflection on experience and represents therefore a second order experience of the world". She proposes that university learning is 'mediated' by symbolic representations which allow different descriptions of the world to those provided by direct experience. Indeed the two forms of knowledge are contrasting and separate. For example in understanding Newton's concept of force Laurillard points out that principles contradictory to experience must be learned. Similar arguments are made by Sandberg and Wielinga (1991).

In business teaching and possibly many other professional areas, however, Laurillard's argument is only partly applicable. Students who do not have experience of the world in which they will eventually work can benefit from learning descriptions of it. Further, even if learning must go beyond description or produce counter intuitive descriptions, it is not clear that embedding learning in real contexts is detrimental. Further, realism may have other advantages such as improved motivation, engagement or memory.

Other critics including Orey and Nelson (1994) suggest the extent to which learning is situated even in schools has been exaggerated. They cite evidence that school learners have both formal and informal (situated) knowledge, and that these are overlapping and interactive. Some understanding of the mix of these different types of learning is important to developers looking to use the IMM literature based on situated cognition. There are a range of learning situations in which first order (experience based) and second order (abstract) learning are of different value. There are perhaps few in which either sort of learning is the whole object of university teaching: however discussion of this issue is based on a very narrowly range of teaching areas at present.

Assumption 5: Realism is a straightforward value free concept: the world is real, computer representations are not

From the perspective of social sciences such as business, it is unfortunate that so much of the IMM literature derives from studies of industry training and teaching about the physical and biological worlds where the issues of representation are may be more straightforward. One problem in the present context is the issue of whether social knowledge is objective or subjective, or more importantly whether computer simulations may lead students to either view through artefacts of their design.

A second and related problem lies in the possibility that mediated realities may become more 'real' than experienced ones. Post-modernist writers and semioticians suggest that it is increasingly hard to determine what is real and what is media derived - indeed the latter is becoming more real than the former, as the title of Baudrillard's (1983) paper "The Precession of Simulacra" is known for highlighting.

In the context of teaching business this idea has the important consequence of reminding designers of their obligations to consider the values and ethical dimensions of created realities. Do students learn that managers are white middle aged males as an unintended consequence of basing a simulation on a real organisation which happens to have such managers? Are there more subtle messages about social relationships, power and organisational values that will be taken by students without work experience as statements of how the world is or should be? Do computer simulations, by implying they describe real worlds, have too much authority?

There are two issues here for developers. One is whether they should consider permitting alternative subjectivities in their designed worlds. The second is whether they should consider inviting students to critique the worlds created, to come to the computer with a mindset that looks not just for descriptions of reality but also for the distorting power of such descriptions.

Progressing the study of realism in social simulations

This paper has attempted to discover guidelines for developers of educational simulations in the literatures of training systems design, simulation and gaming, psychology and education. A major difficulty in each case turns out to be the generalisation of principles to the type of learning sought in 'substitute for experience' simulations in social contexts; most studies are based on learning of either simpler problems (eg learning of geometry or physics concepts; studies of memory for single words) or differently complex, often more deterministic problems (eg aircraft simulators; mathematical models of stock markets).

This problem could be remedied by use of more appropriate stimulus materials in IMM research. Unfortunately this is not likely to happen for two reasons. First technology appears to develop faster than carefully controlled studies can be carried out. For example recent reviews of the role of colour in IMM interfaces lament the fact that most research has used different forms of colour to those now available; research on graphics suffers from being conducted on black and white, C, low resolution screens. If such relatively simple aspects of IMM cannot be understood, what hope is there of knowing how students use realistic representations in learning at the different levels identified above. And secondly controlled scientific studies rarely help developers understand what proportion of the variance in student learning and memory is attributable to the use of either low level media features or higher ones such as the difference between pictures and verbal descriptions, line drawings and photographs, spoken and written text. Two reasons for this are that effect sizes are sacrificed to the quest for significant p values, and that such studies are usually conducted a long way from the natural settings of a classroom.

Ullmer (1994) raises an interestingly parallel problem in medicine: new drugs proliferate faster than researchers' abilities to conduct properly controlled studies in the traditional highly controlled scientific manner. A new approach to testing involves "large simple trials" where large numbers of subjects are tested with very simple measures, for example outcomes classified as hospitalisation, heart attack or death. It would be intriguing to know of interactive multimedia materials that were widely used and known to have strongly helped students learn important concepts, practitioner relevant skills, attitudes, generic skills, meta-cognitive skills, ethical frameworks - anything! Which aspects of these materials most aided that learning could be investigated as a second generation research issue.

Some illustrative research results

An research project currently underway involves looking at how students use a popular IMM product for teaching Total Quality Management at tertiary level, using phenomenographic interviews (Laurillard 1993) and observation. Early results show that despite the primitive technology, involving photographs of coarse resolution and the voices of actors, students saw the simulations as quite realistic. They showed evidence of the powerful constructive processes described by Laurel (1993), building rich internal representations of the offices and people represented. These often included favourable and unfavourable emotions towards the individuals and office environments portrayed.

Most students were also quite enthusiastic about the opportunity to study the application of theory in simulated worlds. Features mentioned included the ability to 'replay' simulations to see additional layers of meaning or recall forgotten aspects. Although students were provided with a linear outline of the contents, observation and self report showed them to work very non-linearly through the material. It was evident they were improving understanding through an active process of moving between different elements, including theoretical principles, graphic illustrations and the simulations. Undoubtedly students do this with textbooks, but the ease with which it can be done in IMM along with the attraction of the simulated rather than verbally described environment suggests considerable scope for deeper understandings of the application of knowledge.

However, enthusiasm for this approach was mediated by students maturity. Some younger students were more inclined to see the simulations as less useful than lectures or tutorials because the latter "forced you to learn what the lecturer wanted". Older students with work experience were more appreciative and could see how the knowledge might transfer to future work situations.

Also to be investigated in this study are self report and 6 month follow up tests of students' memory, including memory for visual and auditory details that might act as retrieval cues for the more abstract subject matter. Interviews completed so far reveal a wide range of individual differences in the extent to which students thought visual images and sounds would help them remember the key points of the package; whether these translate into improved memory is yet to be seen.


"Technology requires you to respect its authority and participate in its workings" (Ullmer 1995), a statement that seems to apply to IMM developers as well as the IMM users that Ullmer had in mind. If developers are not to be driven by the workings of virtual reality technologies they must start a deeper discussion of issues such as those outlined here. New forms of inquiry will be needed to support such discussions.


Brown, J. S., Collins, A. and Duguid, P. (1989). Situated cognition and the culture of learning. Educational Researcher, 18(1), 32-42.

Baudrillard, J. (1983). The precession of simulacra. Art and Text, 11. Reprinted in D. Hlynka and J. C. Belland (Eds), Paradigms regained: The uses of illuminative, semiotic and post-modern criticism as modes of inquiry in educational technology. Engelwood Cliffs: Educational Technology Publications.

Cognition and Technology Group at Vanderbilt (1992). The Jasper experiment: The exploration of issues in learning and instructional design. Educational Technology, 40(1), 65-80.

Dale, J. (1954). Audio-visual methods in teaching. NY: Dryden Press.

Gagne, R. M. (1954). Training devices and simulators: Some research issues. American Psychologist, 9(7), 95-107.

Gagne, R. M. and Glaser (1987). Foundations in learning research. In R. M. Gagne (Ed) Instructional technology: Foundations. Hillsdale, NJ: Erlbaum.

Geiselman, R. E. and Crawley, J. M. (1983). Incidental processing of speaker characteristics: Voice as connotative information. Journal of Verbal Learning and Verbal Behavior, 22, 15-23.

Gibson, J. J. (1986). The ecological approach to visual perception. 2nd Ed. Hillsdale, NJ: Erlbaum.

Goodman, F. L. (1995). Practice in theory. Simulation and Gaming, 178-189.

Hapeshi and Jones (1992). Interactive multimedia for instruction: A cognitive analysis of the role of audition and vision. International Journal of Human-Computer Interaction, 4(1), 79-99.

Hays, R. T. & Springer, M. J. (1988). Simulation fidelity in training system design. NY: Springer-Verlag.

Jacoby, L. L. and Brooks, L. R. (1984). Nonanalytic cognition: Memory, perception and concept learning. The Psychology of Learning and Motivation, 18, 1-47.

Laurel, B. (1993). Computers as theatre. Reading, Mass: Addison-Wesley.

Laurillard, D. (1993). Rethinking university teaching. London: Routledge.

McClelland, J. L.(1994). The organization of memory: A parallel distributed processing perspective. Revue Neurologique, 150(8-9), 570-579.

Orey and Nelson (1994). Situated learning and the limits of applying the results of these data to theories of cognitive apprenticeship. In Proceedings of Selected Research and Development Presentations at the 1994 National Convention of the Association for Educational Communications and Technology, Nashville, TN.

Paivio, A. (1990). Mental representations: A dual coding approach. NY: Oxford University Press.

Reiber, L. (1990). Animation in computer-based instruction. Educational Technology Research and Development, 38(1), 77-86.

Reiber, L. (1995). A historical view of visualization in human cognition. Educational Technology Research and Development, 43(1), 45-56.

Salomon, G. (1978). Interaction of media, cognition and learning. San Francisco: Jossey-Bass.

Sandberg, J. and Wielinga, B. (1991). How situated is cognition? Proceedings of the Twelfth International Conference on Artificial Intelligence, Vol 1. Sydney.

Schlecter, T. M. (1993). Computer based instruction and the practical aspects of memory. Applied Cognitive Psychology, 7(7), 653-665.

Sims, R. J. (1991). Simulation versus emulation and the rules of the game. Proceedings of ASCILITE 91, Geelong.

Ullmer, E. J. (1994). Media and learning: Are there two kinds of truth? Educational Technology Research and Development, 42(1), 21-32.

Young, M. F. (1993). Instructional design for situated learning. Educational Technology Research and Development, 41(1), 43-58.

Author: Dr Peter Standen
Department of Management
Edith Cowan University
Churchlands WA 6018
Ph (09) 383 8335 Fax (09) 273 8754

Please cite as: Standen, P. (1996). Realism and imagination in educational multimedia simulations. In C. McBeath and R. Atkinson (Eds), The Learning Superhighway: New world? New worries? Proceedings of the Third International Interactive Multimedia Symposium, 384-390. Perth, Western Australia, 21-25 January. Promaco Conventions. http://www.aset.org.au/confs/iims/1996/ry/standen.html

[ IIMS 96 contents ] [ IIMS Main ] [ ASET home ]
This URL: http://www.aset.org.au/confs/iims/1996/ry/standen.html
© 1996 Promaco Conventions. Reproduced by permission. Last revision: 15 Jan 2004. Editor: Roger Atkinson
Previous URL 24 Feb 2001 to 30 Sep 2002: http://cleo.murdoch.edu.au/gen/aset/confs/iims/96/ry/standen.html