Designing to Support the Development of Modeling Practices and Epistemologies in Middle School Science

• In Event: 32.029 - Supporting Science as a Modeling Practice in the Classroom Through the Lens of Next Generation Science Standards

Abstract

Objectives. Modeling is epistemically challenging for K-12 students. Typically, school structures do not support students in seeing models as generative, predictive, revisable tools. Instead, students perceive models as answers for teachers. Although existing learning progressions describe students developing modeling practices, the highest levels of the progressions are hypothetical, because few students have demonstrated advanced modeling practices in these contexts (Fortus, Rosenfeld, & Schwartz, 2010; Schwarz et al., 2012). In order to (1) refine the modeling learning progression, (2) operationalize the highest levels of the progression, and (3) identify curriculum and classroom culture supports for developing advanced modeling practices, I conducted a semester-long study of eighth graders engaging in diagrammatic, physical, and computational modeling (Figure 1).
Theoretical Framework. Following Schwarz et al. (2012), I focus on explanatory mechanistic models, which are critical to building science knowledge (Russ, Scherr, Hammer, & Mikeska, 2008). My design and analyses stem from Schwarz, Fortus, and colleagues’ (2012) articulation of learning goals for modeling (Fortus et al. 2010; Schwarz et al., 2009; Schwarz et al., 2010; Schwarz et al., 2012). This progression details five categories of epistemic perspectives that develop as students engage in modeling (Salience and Generality, Audience, Evidence, Mechanistic and Generative, and Revision). I chose this progression because it is one of the most elaborated in K-12 modeling literature, is consistent with the findings of other researchers in this field (Bamberger & Davis, 2013; Berland et al. 2015; Krell, zu Belzen, & Krüger, 2014), and reflects the science-as-a-practice perspective by analyzing students’ actions as evidence of their epistemologies and engagement in practices.
Methods and Data. I conducted an iterative design-based classroom experiment (Cobb, Confrey, diSessa, Lehrer, & Schauble, 2003). I met with the students’ teacher weekly to develop and test conjectures. Data include video of each class period, audio of conversations with students during group and individual work, and artifacts (models, written reflections, interviews, and assessments).
Results. I found students reaching the highest levels of the learning progression across all five categories of epistemic perspectives, enabling me to operationalize the highest level of each category (see Table 1 for an example of student growth in one category). In terms of curricula, creating and interrelating diagrammatic, physical, and computational models supported conceptual growth for students because this perspective-shift encourages students to re-negotiate their conceptions through constraint satisfaction, leading to more coherent understandings (Greeno & van de Sande, 2007, 2012). Moreover, each of the three model types highlights unique modeling practices: diagrammatic models emphasize the descriptive and communicative nature of models (Latour, 1990), physical models encourage students to grapple with the material challenges of designing measures and interpreting data (Penner, Lehrer, & Shauble, 1998), and computational models facilitate abstraction and prediction (Sengupta et al., 2013). In terms of culture, I found that increasing student agency by not assigning grades, limiting authoritative resources, and encouraging student-designed questions and investigations supported learning and engagement.
Scholarly significance. This work addresses the need for knowledge about supporting the development of classroom modeling practices emphasized in the session proposal.

Author

• Ashlyn Karan, Vanderbilt University