Contributors: Chad Dorsey
Computational models and simulations allow students to understand through exploration the behavior of systems that are difficult to understand by other means. Simulations are able to represent systems through a sufficiently close approximation of the true behavior and rules that underly their real-world counterparts. In the process, models and simulations can permit students to observe and investigate the represented scenarios by manipulating parameters at will and observing the effects of their actions. There is a wide range of examples of what can conceivably be termed a simulation. In one extreme, the output of a simulation can be purely mathematical, with the computation resulting only in numerical output that must be interpreted by the student or user. In the other extreme, a simulation can provide rich visual portrayal of the phenomena being modeled and can often provide multiple representations of a scenario.
For the purposes of this definition, we distinguish between “simulations” and “animations” with the distinction hinging on whether the output accessed by a student is predetermined or derived from underlying rules and generated based on those rules. A constructed video or Flash animation may represent a molecular process in a detailed fashion and be very useful for teaching a given concept while not adhering to the definition of a simulation. If the output of the example is the same each time it is run, this is a sign that it is an animation (sometimes referred to as a “dynamic visualization”). Often more importantly, simulations contain parameters that can be adjusted by the user to influence the represented outcomes.
Simulations have powerful capabilities to make exploration of the natural world possible in new ways. Simulations frequently provide a visualization of processes that would otherwise be invisible. In other cases, simulations may permit exploration of concepts that are too big to be readily observed, happen on too long or too short a timescale for practical observation, or are too dangerous or impractical to reproduce in a classroom.
With simulations, students can engage with — and perform inquiry learning about –- concepts that would often be difficult to conceive of otherwise. This has enormous ramifications for learning in STEM fields, especially where learning about the natural world hinges upon a wide variety of concepts that are frequently derived from a small number of unifying ideas and principles. Simulations built upon these core principles can aiding learning by recreating many behaviors that model the natural world.
Simulations have come a long way since the initial widespread existence of personal computers made them a possibility a few decades ago. Simulations for science learning arguably had their roots in the development of modeling as an important force in scientific research. Today, simulations are fundamental to most scientific research, permitting scientists to verify hypotheses and empirical observations by demonstrating how a simulation does or does not reproduce findings or predict verifiable new results. Educational simulations have also evolved. Basic simulations designed decades ago to illustrate a single mathematical concept or display a physical concept in a rudimentary fashion have given way to simulations with complex modeling and multiple manipulable parameters, engaging and immersive visual environments, and linked models that represent multiple aspects of a conceptual domain.
The future of educational simulations is especially bright. With the extension of this evolution, educational technology will enter into a new era of simulation enabling investigation and understanding far beyond current possibilities. Increased computing power has already raised the calculation possibilities of simulations to new heights, a trend that shows no sign of slowing. And experience on the part of both designers and users permits the development of simulations that are both more complex and more usable than ever before.
As these conditions combine, simulations will enable new, very complex understanding for students in ways not currently thought possible. Simulations will be able to model large-scale phenomena such as climate processes or population evolution effectively. Multiple-scale, linked models will permit students to drill down into a concept, modeling it at all levels and linking concepts. For example, a student may experiment with a simulation representing a population of organisms to investigate population genetics. The student becomes interested in one specific individual and investigates further. As she does, she zooms in to see the organism and its chromosomes, able to influence the organism’s individual alleles or cause new mutations in the organism’s DNA. The changes the student makes affect a linked cellular-level model of metabolism in the organism; investigating this, the student can see that the mutations she caused have created specific molecular changes to cause the metabolic effects. The student is then able to zoom back out, breed this organism multiple times, and observe the changes as the mutation makes its way through the overall population.
By developing simulations that are sufficiently complex and interlinked, an additional, very important, factor becomes visible. Simulation environments that are rich enough to be explored in great depth foster behaviors that directly mimic the endeavor of scientific exploration and discovery. A student investigating a simulation environment is, for all extents and purposes, undertaking the process of scientific inquiry. If the simulation is rich enough and scaffolded sufficiently that it enables a student to pose and test new conjectures and includes a community of learners among whom the conjectures can be shared and debated based on evidence from the simulation, the resulting environments can evoke the process of scientific inquiry better than any situation that can be created in the classroom. By engaging in guided inquiry through work with a simulation, students can become in effect practicing scientists within a miniature scientific community. Since many scientific communities today conduct both their investigations and correspondence via virtual means, the practice of exploring simulations by students and the practice of working scientists will rapidly converge in the future. In this vision, students could eventually move seamlessly from work in “practice” simulations to actual current research in scientific domains.
The creation of simulations has flourished with the expansion of tools for programming and visualization. This has resulted in a large number of overlapping examples and some new innovation in simulation design. A major challenge moving forward toward future simulations will be in tackling the creation of large-scale, multi-level simulations that can approximate larger portions of scientific ideas. Doing this properly, with simultaneous attention to veracity and simplicity of presentation, is a challenge that currently lies between daunting and overwhelming but offers a tantalizing promise nonetheless. An additional challenge relates to assessment of learning via simulations. For a variety of reasons, the literature on learning via simulations is somewhat divided as to pedagogical effectiveness. Assessing the effectiveness of simulations in the future and building up this literature will be a challenge. The development of simulations that are able, in themselves, to provide direct data about student explorations and the consequent development of simulations that can be used as performance assessments of science learning are challenges that the field has begun to tackle. These pursuits are still in their infancy for the most part, and bringing them into reality will be a major challenge in the future of simulation development.
Dissenting Views / Critiques
For all their perceived veracity, simulations remain only simulations. All creator and users of simulations would do well to recall the words of statistician George Box that “All models are wrong. Some are useful.” Both creators and users of simulations run the risk of forgetting this regularly and deriving ideas or principles from work with simulations that are invalid. This is reason that simulations need always be presented together with an explanation of their assumptions and limitations. Even with properly presented explanations of these assumptions, simulations involving visualizations run the risk of being “deceptively clear,” as users see the results and think that their understanding of a simplified simulation is a sufficient stand-in for complete understanding of the topic. Because of simulations’ simplicity, depth of understanding and richness of real-world phenomena can easily be lost. Similarly, or simply the importance of noisy additions to the data in the real world is generally kept out of simulations for clarity’s sake. Because simulations take into account only the rules that are understood, they may present a version of physical phenomena that is inherently inaccurate because of this.