Argumentation Using Evidence in a Modeling Task: Challenges in Model-Evidence Coordination

• In Event: Learning to Argue From Evidence

Abstract

Argumentation engages students in constructing and critiquing claims about phenomena using appropriate and sufficient evidence (e.g., McNeill & Krajcik, 2011). How students use evidence in arguments provides insight into their ability to interpret, evaluate, and relate evidence to claims. However, students’ judgments about evidence quality and about appropriate conclusions can be at odds with teachers’ expectations (Manz & Renga, 2017). We know less about the different ways in which students interpret and evaluate evidence, or how interpretations are influenced by the nature of the evidence and the design of instructional tasks. A better understanding of how students reason with and about evidence is critical to understanding argumentation processes and to improving these processes. In this paper, we explore these issues through the examination of student-generated arguments about a genetic disorder part of a written assessment. Our research question is: in a written argumentation task, how do students interpret and evaluate evidence in relation to two explanatory models?

This assessment followed an 8-week model-based inquiry instructional unit in genetics. In the assessment task, we presented 11th grade students (n=294) with two competing models of a genetic skin disorder. One of the models presented the correct mechanism (connectin), while the other provided a compelling but not biologically likely explanation (separatin). Students were provided with six pieces of evidence (Figure 1) of different types including poor quality anecdotal support for the implausible model, summaries of empirical studies supporting the correct model, and diagrams and descriptions of the phenomenon.

We evaluated evidence usage as a function of model selection and focused our analyses on the students (31%) who chose the incorrect, separatin model to better understand these students' reasoning. These students, unsurprisingly, used the anecdotal piece more often than those who chose the correct model. However, some students who chose the incorrect model also used evidence pieces designed to support the other (correct) model. In these cases, students took one of three approaches in relating evidence and models: (a) reinterpreting the conclusion of the empirical pieces and making connections between the empirical piece(s) and the anecdote, (b) providing an alternative plausible interpretation of the evidence in support of the incorrect model, and (c) misunderstanding the model and then interpreting the evidence in plausible ways to support their misconstrual.

In retrospect, there were aspects of the models and evidence that afforded these alternative explanations. Thus, even students who selected the incorrect model provided evidentiary support by interpreting and co-opting evidence in interesting, plausible ways. Such maneuvering is possible because models are inevitably underdetermined by small body of evidences; the evidence will not be completely conclusive. It is critical for educators and education researchers to recognize that there is reasonable flexibility in how students interpret evidence and models. Designers and teachers need to be aware of this when engaging students in evidence-based argumentation, and when interpreting students’ arguments. Instruction can focus on enabling students to learn to think through why some interpretations are more consistent with the full body of evidence and with all model components.

Authors

• Veronica Cavera, Rutgers University - New Brunswick/Piscataway

• Ravit Golan Duncan, Rutgers University

• Clark A. Chinn, Rutgers University