Dissertation Proposal Announcement Ph.D. in Education Program: Ellie Marino “Supporting Student Interpretation of Mathematical Models in Physics: The Role of AI-Based Assessment and Scaffolding of Mathematics Competencies During Data Analysis in High School Physics”
High school students often struggle to apply mathematical competencies during science investigations, particularly when analyzing data to justify their claims. While the Next Generation Science Standards (NGSS) emphasize this integration through Practices 2, 4, and 5, many students rely on surface level strategies without developing a deeper understanding. This dissertation investigates how automated assessment and scaffolding can support students with data analysis and interpretation in high school physics virtual labs completed using Inq-ITS, an intelligent tutoring system.
Three studies will examine student performance during the Analyze Data stage of NGSS aligned virtual investigations. Study 1 establishes baseline performance of analyzing data without automated scaffolding to identify challenges with mathematics that students face in this stage. Study 2 uses a randomized design to test the effects of real-time, AI-based scaffolds on competencies related to analyzing data, such as interpreting model fit and evaluating mathematical relationships underlying phenomena. Study 3 explores how students interact with the same scaffolds on analyzing data and whether prior support improves outcomes on those competencies, thus testing generalizability and transfer.
Informed by evidence centered design, learning analytics, and educational data mining, the research evaluates how scaffolds affect sub-competencies of mathematics needed for student reasoning, operationalized as claim-model alignment and model evaluation. Findings will inform learning theory of how students integrate mathematics with science practices and guide the design of adaptive educational technologies, contributing insights into how to support authentic mathematical reasoning as a foundation for deep science understanding.