| This dissertation investigates students' construction of a functional understanding of a Newtonian microworld called the "Envisioning Machine." Prior research highlights the importance of developing students' qualitative understanding through their own constructive learning, and demonstrates the potentials of computer simulations in this regard. A primary concern here is to understanding how students develop scientific understanding through incremental transformation of their knowledge.;The inquiry focuses on three questions: (1) What sorts of knowledge do students' construct? (2) How does their knowledge compare to scientists' knowledge? (3) What role does prior knowledge play in their learning?;My thesis is that students' understanding develops through their construction of a knowledge system comprising registrations, qualitative cases, and p-prims. By "registrations," I mean features that students perceive, label, and select for attention. By "qualitative cases," I mean case-specific schemata containing heuristic associations useful for qualitative problem solving. By "p-prim," I mean physical metaphors drawn from everyday experience that serve as building blocks for explanations.;Students' learning in the Envisioning Machine microworld was videotaped. The resulting student performances were extensively analyzed to produce (a) case studies of learning, (b) models of problem solving, and (c) records of trends in explanations. Based on these observations and additional interviews and tests, I argue that the three forms of constructed knowledge listed above account for students' growing ability to solve qualitative physics problems and give qualitative explanations.;Superficially, these forms of knowledge show little similarity to textbook science, which suggests a negative pedagogical outcome. However, I show that students gradually achieved a good approximation to scientific understanding by reformulating their registrations, qualitative cases, and p-prims. Nonetheless, the data shows definite limits in the degree of approximation achieved. Although all students progressed in problem solving ability, some did not produce explanations consistent with the scientific definitions of velocity and acceleration. Moreover, the properties of generality, integration, and stability, important to scientists, were problematic for students to attain. Despite these difficulties, the demonstration that students can construct knowledge that approximates scientific understanding is an advance. It offers a basis for building more accurate models of students' knowledge and learning, and for reformulating science pedagogy to take advantage of continuities between students' and scientists' knowledge. |
