Developing A Next Generation Concept Inventory Using Methods from Artificial Intelligence

Author(s):
Amanda Manzanares
Graduate Research Assistant
University of Northern Colorado

Need: Unlike more established and single disciplinary fields, Interdisciplinary Environmental Programs (IEPs) comprise a complex integration of social and natural sciences centered on coupled human-nature systems and sustainable natural resource management. Due to an undefined identity, the inherent interdisciplinary nature of these IEPs does not lend itself to traditional discipline-specific concept inventory frameworks for critically evaluating students’ alternative conceptions, prior knowledge, and learning. However, traditional discipline-specific concept inventory frameworks often fall short in adequately assessing alternative conceptions and facilitating learning within these programs. As a result, there’s a pressing need for innovative approaches tailored to the unique complexities of interdisciplinary environmental studies. One example of concept integration is the Food-Energy-Water (FEW) Nexus, which connects environmental processes, management, policy, and socioeconomics of FEW resources. There is a need for education research and collaboration in the FEW Nexus, as evidenced by the recently funded National Collaborative for Research on Food, Energy, and Water Education (NC-FEW).Guiding Question: This four-year project investigated the development of a tool for rapid and large-scale assessment of short answer responses in college-level IEP courses. The overall aim was to develop and test a Next Generation Concept Inventory (NGCI) assessment tool that includes interdisciplinary, systems-level concepts related to the FEW Nexus using constructed response (CR) assessment items. In order to accomplish this, we collected course materials from a range of IEP courses, investigated student thinking on FEW topics, designed and piloted assessment items on key FEW topics and engaged in machine learning model development. Outcomes: Presented outcomes from this work include (1) assessment of FEW concepts covered in introductory IEP courses, (2) assessment of student knowledge of FEW and Climate Change interactions, (3) an exploration of the use of drawing for assessing student knowledge, (4) presentation of a process for using machine learning to assess how students connect FEW concepts, and (5) the use of FEW Nexus as a framework for meeting sustainability education goals. Upcoming outcomes include (1) process validation for assessment tool development for complex environmental concepts, (2) recommendations for improving IEPs curriculum and (3) evaluation of the relationship between understanding of foundational concepts related to complex, integrated concepts in IEP courses. Broader Impacts: First, we anticipate that the information gleaned from reviewing environmental curricula across the United States, combined with CI results showing student learning, will better inform those making curricular and staffing decisions regarding college IEPs. Second, the development of the NGCI containing a set of constructed-response assessment items focused on the complex systems will provide faculty with a valid and reliable instrument for evaluating learning within their courses and programs. This advancement is expected to enhance student learning outcomes, offer valuable guidance for updating current courses and creating new ones, and concurrently contribute to the science education research community by disseminating research on an innovative approach to constructing CI tests and undergraduate thinking about critical FEW topics.

Coauthors

Caterina Belle Azzarello, University of Northern Colorado, Greeley, CO; Kevin C. Haudek CREATE for STEM Institute, Michigan State University, East Lansing, MI; Heqiao Wang CREATE for STEM Institute, Michigan State University, East Lansing, MI; Emily A Royse, Aims Community College, Greeley, CO; Lydia Horne, Unity Environmental University, Distance Learning, New Gloucester, ME; Shirley Vincent, Vincent Evaluation Consulting LLC, Tulsa, OK; Steven W. Anderson Department of Earth and Atmospheric Sciences, University of Northern Colorado, 501 20th St., Greeley, CO; Chelsie L. Romulo, University of Northern Colorado, Greeley, CO