Home Page About CC-ATOMS Proposal Learning from Models

Proposal Contents: Summary Introduction Need Goals Research Activities

INTRODUCTION

To understand much of modern technology, students need robust mental models of atoms, molecules, and their interactions. Many of the areas of greatest job growth, such as biotechnology, medical technologies, nursing and allied health occupations such as nutrition, microcircuits, and photonics, depend on exploiting the properties of atoms and molecules. Almost half of the fastest growing jobs (13 out of 30), for example, are in biological sciences and health related fields (CT Dept. of Labor, 2001). To prepare students to understand future technologies in areas such as these, they need to be conversant with atomic-scale interactions and how these determine macroscopic phenomena. This is particularly true as the chemical basis of living systems becomes better understood and more important in areas as diverse as law, forensics, farming, and nano-engineering.

Learning from Models
There is considerable evidence that students have major misconceptions about atoms and molecules and few ideas about how the forces and motions at this scale relate to macroscopic properties. Even though they are familiar with the idea of atoms, many students believe that there is some continuous matter between atoms and that the properties of atoms such as color, strength, and stickiness, are simply identical to macroscopic properties (Anderson, 1990; Lee, et al 1993; Nussbaum 1985). For instance, students believe that the molecules making up paper would catch fire and that molecules of ice are cold. Furthermore, few students are aware of electrostatic forces and their dominant role at the atomic scale (Anderson, 1990; Children's Learning in Science, 1987).

Initial studies have demonstrated the educational value of highly interactive research-based models of atomic and molecular systems. (Hakerem, 1994; 1996; Birk, 1997; Tinker, 2001a) Surprisingly complex models of atoms and molecules can now be run on the inexpensive desktop computers that are widely available to students. From a few basic potentials and rules, these models can compute the motion of a few hundred atoms while simultaneously displaying a representation of the motion of the ensemble. This can be done quickly enough for students to see the motion and explore the effects of various parameters on the evolution of many systems.

Using such models, we can demonstrate the phases of matter, the gas laws, phase change, latent heat, diffusion, gas absorption, osmosis, thermal diffusion, conformational changes, and the properties of liquid crystals. Student explorations of these models lead to a good understanding of connections between atomic-scale events and what they observe at the macroscopic scale. Using these models and supporting curricula, students are able to predict new micro-macro connections that they have not previously encountered (Tinker, 2001c).

These dynamic, interactive models create very powerful learning experiences. It is hard for instance, to understand osmosis from a written or verbal description of diffusion through a semi-permeable membrane, but it is almost obvious when you see a dynamic model of it actually happening. The learning experience is particularly effective because the model is more than an animation and students can experiment with the system to find out how it "works", that is, what combinations of molecule size, density, pore size, temperature contribute to osmotic pressure. This creates the kind of rich, highly associated mental network of concepts that we know contributes to lasting understanding (Snir and Smith,1995; Jackson, et al, 1994).

Models of Atomic-Scale Interactions
Our work to date has concentrated on models that explain phenomena encountered primarily in physics and physical chemistry. This has resulted in the Molecular Workbench model currently being tested with students. To exploit the educational value of our modeling-based approach in areas that depend on chemistry, biology, and biochemistry, we need to extend the Molecular Workbench model to include covalent bonds, photon interactions, conformation, and other functions. These will permit students to explore a range of new phenomena such as chemical reactions, equilibria, reaction rates, catalysis, molecular form and function, color, spectra, phosphorescence, and fluorescence, many in three dimensions.

These additions to Molecular Workbench will require major technical advances in computational models. There are no general molecular dynamics models that include chemical reactions and light-photon interactions. While creating such models sounds ambitious, the reason that these additions are feasible in this project is that we are not building a research model; we can judiciously trade off precision for educational value. For instance, we do not need to model a quantum mechanical solution of covalent bonds. We have already created a heuristic model of monovalent reactions that conserves energy. This is sufficiently rich to model, for instance, the role of energy and entropy differences in shifting chemical equilibria, a set of important ideas that most students currently must take on faith. We plan to incorporate similar heuristic rules for multi-valent bonding and photon interactions that support the development of accurate student concepts while avoiding some of the confusing complexity of more complete solutions.

Figure 1. A Molecular Workbench model of osmosis. The vertical row of atoms in the center that make up the membrane are tied to the background with harmonic constraints.

Pedagogical Support
A sophisticated model is not, by itself, much use in education. If students are asked to learn by exploring such models, they generally become lost and bored. If they bravely persist, they may discover some of the underlying science, but will likely miss critical parts and fail to transfer their knowledge to other situations or to tests. The model needs to be incorporated into instructional activities that employ effective learning strategies (Bent, 1984; Buckley, 2000; Harrison & Treagust, 1996; Horwitz, 1995-6; Horwitz & Christie 1999; Lee, et al, 1993: Snir and Smith, 1995).
Computer-based learning environments can make inquiry-based learning more successful by offering students cognitive and procedural guidance and by freeing teachers to interact with students about complex science topics (diSessa, 1992; Vanderbilt, 1997; Feurzeig & Roberts, 1999; Slotta & Linn, 1999; Linn & Hsi, 2000; White & Frederickson, 1998). This is important because the best learning strategy for use with complex models is to provide opportunities for student learning through guided inquiry in an engaging social context (Glaser, 1976; Champagne, et al, 1980; Vanderbilt, 1997; Driver, 1985; Collins, et al, 1991; Scardamalia & Bereiter, 1992; Brown & Campione, 1994; White & Frederickson, 1998). The guidance typically involves giving students a question or challenge, providing a simplified version of the model to explore, asking for predictions, probing for reflection, and providing help if the student is confused. Guided exploration of this kind converts a model into a learning activity.

The proposing partners have developed two complementary technologies­Pedagogica and WISE­that support inquiry-based guided explorations of models. Pedagogica is an application that allows curriculum developers to write scripts that can make many specific scaffolded models called "hypermodels" (Horwitz & Tinker, 2001; Tinker, 2001d). Many scaffolded hypermodels can be derived from one complex model such as the Molecular Workbench. Individual hypermodels provide help in exploring and understanding a specific aspect of a model.

For example, as mentioned above, one of the many emergent phenomena that Molecular Workbench software can model is osmosis. Figure 1 shows Molecular Workbench software configured for explorations of osmosis. However, as part of a lesson, students may need to focus on just one aspect of the system, such as the relative sizes of the molecules. Pedagogica can be used to create a hypermodel for such a lesson with only the needed controls, supporting text, and a coordinated macroscopic model, as shown in Figure 2.

WISE, as illustrated in Figure 3, provides a friendly, teacher-controlled server-side environment for combining hypermodels with other online resources into complete instructional modules that can be delivered entirely on the Web (Slotta & Linn, 2000). Together, Pedagogica and WISE create a powerful learning and distribution system that exploits and delivers the educational value of models such as those developed in the Molecular Workbench project.

This material is based upon work supported by the National Science Foundation under Grant No. EIA-0219345. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.