Supporting Evidence and Reasoning Through Data Collected From Scientific Modeling Tasks in an Elementary Science Classroom

• In Event: A New Generation of Goals for Scientific Modeling: Strategies and Lessons Learned for Reach and Impact

Abstract

Purpose/Significance. Scientific simulations can make complex concepts accessible and inviting to students as they learn to generate hypotheses, collect evidence from data, and construct arguments. However, in previous research, we investigated whether and how students make accurate inferences from data and found that challenges exist in the reasoning process, i.e., connecting claims to the evidence (Author, 2018). Studies have shown that students are seldom supported to engage in reasoning using evidence (e.g., Manz, 2016) in science classrooms. In this study we worked with students in data-rich investigations afforded by the StarLogo Nova tool in order to support student reasoning processes. Students used a number of reflective strategies that emerged from the simulation use that ultimately enabled them to negotiate their scientific understanding.

Theoretical Framework. The use of data as evidence in a thoughtful manner provides opportunities for meaningful inquiry because such practices require the ability to understand and use data effectively to inform decisions (Lee & Wilkerson, 2018). In incorporating data-rich investigations into curricula, epistemic cognition is enacted in the process of decision making (Greene & Yu, 2016). Epistemic cognition is the process of evaluating how knowledge is constructed and used in order to make accurate predictions about the world (Barzilai & Chinn, 2018; Sandoval et al., 2014). The following research question guided this study: To what extent and how do students reason with data as evidence to make accurate inferences of the scientific phenomenon?

Methods. We addressed this question with data collected from a 4th grade science unit on Ecosystems: Plants and Precipitation. Students were asked to make predictions about the relationship between plant growth and precipitation amounts when designing a garden. Students were instructed to continuously revise their ideas based on the data that they collected from running the StarLogo simulation and provide explanations about how their data explained the simulations behavior. We collected two data sources in order to understand the instructional variables that enabled accurate scientific reasoning to emerge: teacher interviews and classroom videos.

Findings/Conclusions. To communicate about and infer meaning from the data, students used different types of strategies such as referring to the observations of previous simulation activities, basing predictions on math, and reading the graph represented from interactions among the variables in the simulation. Students also synthesized plant survivorship data from multiple simulation runs to construct explanations and make inferences regarding water requirements for certain plants. Students were emotionally engaged in running the model and became excited when their predictions on plant survivorship closely matched with the modeling output. Through continual data collection and pattern comparison, teachers reported that students became increasingly better able to make predictions about the health of their garden. Additionally, teachers discussed the functionality of the computational modeling tool to support students in developing reasoning skills as they aligned their emerging data to scientific claims.

Authors

• Jooeun Shim, University of Pennsylvania

• Susan A. Yoon, University of Pennsylvania

• Amanda Cottone, University of Pennsylvania

• Bob Coulter, Missouri Botanical Garden

• Stacey Carman, Missouri Botanical Garden