Author(s):
Need: This project aims to serve the national need for highly effective undergraduate STEM education. Specifically, the project aims to enhance the teaching of rock identification and categorization in introductory undergraduate geoscience classes. Helping students learn about categorization, such as rock types in geology, is important in science curricula because categories are the building blocks of basic thought processes in humans. Categories provide an efficient means for humans to reason and draw inferences about the nature of the world. Guiding Question: To what degree do students learn rock identification from a) manipulating and studying physical rock samples, b) manipulating and studying high quality virtual (3D) models displayable on web-connected laptop computers, and c) manipulating and studying both physical rock samples and high quality virtual (3D) models displayable on web-connected laptop computers? To address the question, we are conducting a classroom-based experiment in which students receive rock-identification training under each of these three conditions, and then engage in a final test phase that assesses their ability to identify novel physical samples from the trained rock categories.Outcomes (key findings and deliverables): We created a set of 116 high quality virtual (3D) rock models, representing 10 common types of each of the major divisions of igneous, sedimentary, and metamorphic rocks. We also created Rockviewer, a web-based, client-server system, developed with a gaming-based engine (Unity). Rockviewer delivers virtual rock models to student laptops and records student interactions with the virtual rocks. Rockviewer is designed as an installable system under Windows. The Rockviewer system was used in the experimental conditions described in the Guiding Question section that involved rock-identification training using the virtual 3D models. Due to the pandemic and a delay of the return of students to campus, we have tested only a single cohort of students in the classroom experiment thus far. The results are going in the hypothesized direction, with students trained using the virtual 3D models performing at least as well during the final test phase as those students trained using actual physical samples. Tests of additional student cohorts are in progress and planned for future semesters. Broader Impacts from proposal: The domain of rock classification appears to be both a challenging and highly representative example of natural-science category learning. Therefore, the experimental results obtained in this project should provide training guidance across the broad expanse of STEM disciplines in which students are tasked with learning and inducing categories from the study of individual objects. The results of the project have the potential to inform and improve standard practice in teaching science categories, and by so doing assist societal efforts to promote STEM education. Finally, the technology that will be used in the project will provide the geoscience-education community with a channel for instruction of students who are currently underserved because they have limited access to physical laboratory facilities.
Coauthors
Scott Brande, UAB, Birmingham, AL; Robert Nosofsky, Indiana University, Bloomington, IN
