Artur Pokropek

Abstract

This talk traces how process data have moved educational measurement beyond the final response toward the behavioral traces left during test-taking. Following a progression of increasingly fine-grained signals, from response times and full action logs to mouse-cursor trajectories and, most recently, webcam-based attention and gaze tracking, I will show how each can be modeled to detect disengagement, guessing, and careless responding, and to better understand how students interact with tasks. I will close with the opportunities and the methodological and ethical challenges that arise as assessment moves toward richer, AI-driven process data.