Have you ever wondered if we might have chosen wrong categories when we divide the world of instruction into classes like tutorial, simulation, job-aid (EPSS), and so forth? As I look beneath the surface of these categories, I am starting to think that we can do better.
We tend to classify instructional products by their mechanism, delivery mode, or experiential features instead of by their instructional essence. "Tutorials" project direct instructional messages and related interactions. "Job-aids" are used to support performance in real environments. "Simulation" encompasses those kinds of product or experience that resemble reality in some way. These definitions miss the instructional nuance and create classes of instructional tools that are huge, have enormous range and variation, and whose names tell you almost nothing about the instructional charateristics or use of the products.
I suggest the term "model-centered instruction" as a means of describing how simulations, as well as a host of things we don't normally call simulations, are used instructionally in ways that cut across the traditional categories. The central principle of model-centered instruction is that the pursuit of the instructional designer is to: (1) provide real systems or models (of three types: environment, system model, and expert model), (2) with which the learner can interact, (3) during the solution of a problem. Problems are assumed to be selected and ordered by principles which maximize learner benefit.
The traditional notion of simulation is included within the scope of this definition, but this is more: it also encompasses real environments, systems, and expert performances and relates them (and simulation experiences) to their specific application during instruction. This definition of an instructional type defines things in terms of how they are used in the instructional act. Not a definition of a product type, it defines a specific context and interaction with the learner--a type and structure of experience.
The heresy gets even better. I would like to propose the idea that all instruction is model-centered. Consider the instructor teaching middle school learners about atoms. Since atoms are not visible to the naked eye, the instructor searches for a representation of some sort that is. What are the options? A computerized model? Virtual? 3-D? 2-D? Animated? Non-computerized? Videotaped? A rapid sequence of photo stills? A handful of isolated stills? And what is the ultimate endpoint of this denaturing of the real into lower-grade representations? A verbal description. If the instructor can find no other means of representing the atom, he or she will resort to verbal explanations of the sort that are common to textbook passages and lectures.
But if you listen to the content of the text or the lecture what do you hear? You hear the description of a system--a verbal model of a system deconstructed by the instructor into verbal representations because that is all that is available. It is not a question of whether we employ model-centered instruction: all instruction is model-centered. It is a question of how far we denature real systems and phenomena into models before we represent them to the learner. The media are simply tools for denaturing real systems and phenomena in order to represent them to the learner.
Model-centered instruction is a useful and important perspective because humans learn and think in terms of (mental) models, not in terms of isolated facts and dissociated elements of knowledge. In the absence of formal instruction, individuals seek experience with real or modeled systems as a source of learning.
What's the impact of this perspective? If it is a correct perspective, it has implications for the way we design instructional experiences, the methods we use to construct them, and the tools we use in doing so. The primary questions of the instructional designer should be, for this learning: (1) what is the appropriate model (or real system) the learner should experience, (2) what is the appropriate level of denaturing for this learner, (3) what sequence of problems should the learner solve with respect to this model or real system, (4) what resources should be available as solving takes place, and (5) what instructional functions to augment the learner's own knowledge and skill should accompany solving?
The method of design changes from analyzing tasks to analyzing environments, systems within environments, and expert behavior toward both. From this analysis comes the specification of a number of problems which the designer may select and sequence. From them comes the specification of the learning environments required, the instructional functionalities required, the surface dramatic dynamic required, and the logical structures required (if a computer is being used). The tools of most value to designers are those which can be used to build models of things for learners to interact with--models with a range of denaturings. This means first and foremost that those tools which we have been calling "simulation-building tools" and treating as a luxury become of primary importance and other representation systems (media) follow in their trail.