Knowledge Integration as a Lens for Structuring and Evaluating Students' Exploratory Practices With Virtual Models

• In Event: Knowledge Integration: Trajectories, Opportunities, and Future Directions

Abstract

Theoretical Framework. Inquiry activities designed from the knowledge integration (KI) perspective prioritize open investigation of models as an opportunity for students to generate and test their ideas (Linn & Eylon, 2006, 2011). However, not all models support the KI processes of expressing students’ own ideas, seeking new ideas, distinguishing ideas, and forming syntheses. Evaluating students’ exploratory strategies when using a model, therefore, provides evidence to support redesign to reinforce KI.
Objectives: In this study, we evaluate process data, automatically collected in the Web-based Inquiry Science Environment (WISE), to evaluate interactions with a model of density and buoyancy, where students typically have persistent, but incomplete ideas (Penner & Klahr, 1996). In this curriculum, students use models to critique incomplete ideas attributed to fictional students (Dan: big, heavy objects float; Aida: heavy objects float, size doesn’t matter). To track student use of the KI processes we use logs of student interactions and written critiques. Finally, as a proximal learning measure, we evaluate a graphing item, presented after the models, in which students plot points to represent sinking and floating objects.
Methods. In our first design iteration students were asked to evaluate Dan and Aida’s claims, depending on randomly assigned conditions, with either a graph or a table to represent their data. Additionally, Dan and Aida’s claims were supported by incorrect graphs or tables that could be modified to explore how students identify discrepant data (Figure 1).
Data sources and Results. Results (Table 1) show that students, in both conditions, had more difficulty refuting Aida’s claim than Dan’s. Classroom observation revealed that students initially struggled to generate appropriate data to refute Aida. Instead, they often produced evidence to support her claim (which often was the same as their claim). Working closely with the classroom teacher, we developed discussion activities to model gathering sufficient evidence. Yet, only about half of students who gathered sufficient evidence explicitly refuted Aida’s claim. Furthermore, few students expressed normative ideas about density. These observations suggest that critiquing claims rarely persuaded students to add new ideas, distinguish these ideas from their own, and come to new conclusions. In part, this may be due to the limitations of the slider interface, which does not enable precise control of mass, but also because of limitations to the critique structure.
Based on this evidence, we redesigned the virtual model to include a new activity that prompted students to express and justify their own conjecture about buoyancy before testing the claims of Dan and Aida. We tested two conditions for making conjectures: draw a graph or fill in a table. Results show that the opportunity to explore their own ideas substantially increased the number of students who made a density claim. Furthermore, students with the table were more likely to discover density while exploring Aida’s claim. Students with the graph were more likely to refute her claim. This divergence between conditions provides opportunities for investigation in future studies (e.g., testing mixed formats).
Significance: Overall, these results indicate that evaluating how KI is enacted in logs of students’ exploratory strategies can strengthen iterative design.

Authors

• Jonathan Michael Vitale, University of California - Berkeley

• Marcia Linn, University of California - Berkeley