Purple Perils redux: Open-ended, AI-resistant reasoning problems for introductory undergraduate sensation and perception instruction

Daniel J. GrahamDepartment of Psychological Science,Hobart & William Smith Colleges,Geneva,NY,USA

DOI: https://doi.org/10.1080/13506285.2023.2250518

2024-01-13

Visual Cognition

Abstract:In terms of pedagogy, there is a need in introductory sensation and perception courses for practice problems that promote critical thinking and elaborate on course material, like those given in the physical sciences. However, perceptual psychology mostly lacks the mathematical foundations that provide ready material for such elaborative assignments. Here I introduce a series of short assignments dubbed Purple Perils, which are inspired by the perception-related missives made famous by J. J. Gibson. These assignments provide students with open-ended opportunities to think through compelling and relatable problems in perception, often with an applied component. By their nature, such assignments are at present somewhat resistant to large language model AI assistance: probes of ChatGPT suggest that students who illicitly use such systems may not gain the synthesis of ideas that Purple Perils demand, and will therefore be disincentivized to use such illicit aid. This kind of assignment can reinforce learning in topics across sensory modalities, and can be used for exam questions. Purple Perils promote critical thinking and quantitative reasoning, e.g., through graphical literacy; reinforce biological knowledge; encourage classroom debate and discussion; and connect to historical and contemporary debates and findings. Gibson's own Purple Peril about the El Greco fallacy leads the series of assignments. A protected, crowd-sourced, repository of Purple Perils is also deployed to allow sharing of questions among instructors.

psychology, experimental