Visualization tools: Models, representations and knowledge integration.

다국어 초록 (Multilingual Abstract)

Learning science requires students to make inferences and difficult to imagine. Scientific visualization is one way to make science and scientific thinking more visible to students. This dissertation investigates how visualization can be utilized for science education by studying how students integrate information from visualizations into their thinking.
For this study, I developed a series of computer visualizations depicting thermodynamic phenomena. Thermodynamics is a topic that is both fundamental for several branches of science and difficult for many students to master (Linn & Songer, 1991). The design of the visualizations was learner centered. Pilot studies suggested that a dot-density representation of temperature would present a visual analogy of temperature as a measure of heat energy density. Energy density is a powerful model that can help students explain everyday heating and cooling phenomena.
Dot-density computer visualizations were introduced into a public middle school science class studying thermodynamics (N = 178). Half of the students used the visualizations, while the other half served as a control. Interviews, classwork and tests were collected from the students in order to determine how the visualizations affected students' learning. Although there were not significant differences in the posttests for the groups, the classwork during the semester showed that the visualizations did affect how students envisioned heat and temperature. Students could often apply the energy density model in their reasoning during visualization activities, but when the visualizations were unavailable, many students applied less useful models.
The interviews illustrated several difficulties that students had in learning from the visualizations. Some students interpreted the visualizations to support their existing conceptions of heat. Other students needed to have a visualization present to cue the energy-density model during problem solving. On the posttest, some students drew images with dots in them, but they lacked the model that underlies the representation. Students who avoided these problems and integrated the visualizations into their thinking were highly successful on the posttest.
These results suggest that for visualizations to be effective learning tools, students need to understand the visualizations and also explore underlying scientific model. Students in this study who connected the visualizations to other ideas about thermodynamics developed a robust understanding of the science. These findings inform our understanding of the science learning process. Students appear to draw from a repertoire of models in their reasoning. Visualizations are a powerful way to introduce models to students, but work best with opportunities for students to integrate the models into their thinking.