The recent interest in having science instruction focus on helping students to construct and evaluate abstract models of situations fits with the recognition that effective science learning calls for students to construct new representations that differ in important ways from those used in everyday life. Science involves more than gathering new data and making inductive generalizations from those data; it also involves new ways of seeing those data in terms of idealized representations. Although there are many approaches to building these models in different domains, science commonly incorporates mathematical relations in these models (Nersessian, 1992) as well as physical intuitions and sensorimotor schemas (Brown, 1993; Clement, 1991). As Nersessian (1989) points out, “in learning Newtonian mechanics, students must usually also learn how to construct an abstract, mathematical representation of the physical world for the first time.” Thus, science educators should not neglect teaching students some of the idealization techniques (such as thought experiments and limiting case analyses) that are central to constructing those abstract representations and that can facilitate their recognition of deep analogies between superficially different phenomena.

Strengthening New Systems of Ideas

Constructing a new system of ideas does not, of course, ensure that these ideas will be internalized (i.e., frequently used in appropriate contexts or that they will even be preferred to one’s initial ways of thinking). How does a new conceptual system become strengthened and gain ascendancy over one’s initial ideas? Many conceptual change researchers have considered that engaging in argument may be a central part of this process (e.g., Chinn and Brewer, 1993; Strike and Posner, 1985; Thagard, 1992). More specifically, students are asked to evaluate (or debate) the adequacy of the new system with known competitors. For example, the new system will gain ascendancy if seen as more plausible (consistent with prior knowledge and existing data) and fruitful (generative of further questions) (Strike and Posner, 1985). Or the new system will be favored if it is seen as more explanatorily coherent (Thagard, 1992); a variety of aspects contribute to judgments of coherence, such as explanatory breadth, elegance, simplicity (not ad hoc), avoidance of contradiction, and future prospects. Even elementary school students are sensitive to many of these features in judging rival accounts. More specifically, Samarapungavan (1992) found that children prefer accounts that explain more, are not ad hoc, are internally consistent, and fit the empirical data. An important step in evaluating an argument may first be to discuss and construct some shared norms for argumentation not only among students but also with the broader scientific community they are trying to understand (Brown and Campione, 1994; Beeth, 1998; Beeth and Hewson, 1999; Duschl and Osborne, 2002; Sandoval, Reiser, 2004).

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement