Emphasizing Evidence Evaluation in Model-Based Inquiry

• In Event: Promoting Epistemic Practices in Science Classes

Abstract

Objectives and theoretical framework. The Next Generation Science Standards represent a shift toward modeling and argumentation as central features of science classes (NGSS, 2013). Recent theoretical work on learning progressions for scientific argumentation (Berland & McNeill, 2010) and modeling (Schwarz et al., 2009) has described levels of sophistication in student reasoning as they develop proficiency with these practices. Both of these useful frameworks, as well as other research, focus on evidence-claim relationships; much less attention has been given to relationships among pieces of evidence. We argue that attention to evidence relations should be a core component of the epistemic practices of modeling and argumentation in the classroom. Here we present findings of middle-school students’ reasoning about multiple pieces of evidence and their relationships (i.e. the body of evidence). We show how carefully designed instruction and scaffolding can support such reasoning.
Data sources and methods. In our PRACCIS (Promoting Reasoning And Conceptual Change In Science) project, we have developed and adapted a number of scaffolds that engage students in key aspects of modeling and argumentation. For example: (a) the Model-Evidence Link Matrix adapted from Suthers and Hundhausen (2003) supports the coordination of evidence and competing models; (b) criteria lists explicate norms about what makes for a good model or argument; and (c) evidence rating tools engage students with evaluating evidence quality. We implemented these scaffolds in a study spanning over more than 4 months of instruction to promote a deeper grasp of evidence among students.
Our analyses focus on two hundred middle school students taught by three teachers in a model-based inquiry unit in genetics. In this unit students evaluated two competing models of the molecular basis for inherited HIV resistance in humans using multiple pieces of evidence. We used mixed methods to analyze students’ ratings of the evidence, how they connected the evidence to both models, and their arguments for the best model.
Results. We identified several patterns of within evidence and evidence-to-model relations identified. A particularly important form of reasoning that emerged in classrooms occurred when students initially dismissed a piece of evidence as being of low quality or irrelevant, only later to change their view of this evidence in light of another piece of “new” evidence. Students would, at times, select one model based on two pieces of evidence initially provided, and then reverse their choice to the competing model after reviewing the entire evidence set after considering the relations among the evidence in the full evidence set. We will discuss the ways in which the lesson design and scaffolds promoted, for at least some students, consideration of the evidence set as a whole.
Significance. Scaffolding students’ thinking about relationships within bodies of evidence (as opposed to thinking about each piece of evidence in isolation) could be a productive step forward in developing theories about what constitutes increasing student sophistication in scientific modeling and argumentation. It is a type of reasoning that has received little attention to date.

Authors

• Ron Rinehart, Rutgers University

• Ravit Golan Duncan, Rutgers University

• Clark A. Chinn, Rutgers University