CREATE for STEM Presentation – “AI for Smarter Science Assessment”

This presentation demonstrates how AI can revolutionize the scoring and interpretation of student data from science assessments through two studies. Study 1 investigates automated scoring of object-based models depicting particle behaviors and macroscopic properties to assess student understanding of the structure and properties of matter. By extracting key features from the models and correlating them with human-assigned scores, the study demonstrates the potential for AI to provide interpretable insights into student thought processes along a learning progression. Study 2 focuses on identifying reasoning patterns related to the three dimensions of the Next Generation Science Standards (NGSS). Using machine learning, we developed an AI-powered tool that classifies reasoning patterns based on student responses to an Ecosystem assessment task. Together, these studies highlight AI’s potential to enhance the scalability and insightfulness of science assessments. 

Light refreshments will be provided for in-person attendees 

Register for Zoom here.

Speaker:

Lei Liu is the Research Director of the K–12 research team at ETS Research & Development and an adjunct professor at the University of Pennsylvania. She holds a PhD in Educational Psychology with a specialization in Learning Sciences and Educational Technology from Rutgers University. Liu’s research spans science education, assessment, and the integration of educational technology. She has led numerous federally funded projects aimed at advancing STEM education through innovations such as learning progressions, AI-supported assessments, and virtual laboratories. In addition to her research, Liu is instrumental in managing large-scale operational programs at ETS, including the California state assessment initiatives and the National Assessment of Educational Progress (NAEP) in science.