Perceptual Training to Combat Visual Misinformation
Misleading graphs can quickly convey false information, posing a severe risk to society. While existing interventions target conceptual processes, visual misinformation often targets perceptual processes, which are automatic and prone to biases. Visual misinformation often targets perceptual over conceptual processing. This project will develop a perceptual training method that will teach viewers to extract correct information from misleading graphs so that they become “immune” against visual misinformation. Funded by NSF IIS 2202457 (RETTL).

Internal Visualization Skills in Engineering
In many STEM domains, such as engineering, visuals are presented to introduce concepts. Later, the visuals are faded out and the concepts are represented mainly through formulas and other symbolic representations, with the expectation that students mentally visualize these concepts. This project investigates how to use visuals in early instruction in a way that helps students internally visualize the concepts later on. Lessons learned will be implemented in an intelligent tutoring system: Signals Tutor. Funded by NSF DUE 1933078 (IUSE).

Intelligent Blending of Physical and Virtual Representations
Physical and virtual representations have complementary benefits. Effective combinations of these representation modes depend on topic-specific conceptual and embodied knowledge. Hence, effective combinations are likely too complex for instructors to achieve without support. We investigate how physical and virtual representations complement one another. Results will inform the development of a an adaptive educational technology that intelligently blends physical and virtual representations. Funded by NSF IIS 1651781 (CAREER).

Use of Representations in Early Childhood
Young children can use representations to explain their thinking about mathematical ideas. However, little is known about how children's ability to navigate multiple types of representations offers insights into their mathematical understanding. To address this gap, this project will interviews teachers and students in 4K classrooms about the use of representations in early math learning. Funded by CRECE.

Collaborative Representations
Learning with visual representations is socially mediated: students often collaboratively make sense of visual representations. We investigate how collaboration can enhance the effectiveness of representation support offered by educational technologies, while also using educational technologies to support effective collaboration. Studies in chemistry courses test how to use the cognitive model of Chem Tutor to identify individual misconceptions and to prompt students to collaboratively discuss mistakes resulting from these misconceptions.

Adaptive Support for Connection Making among Multiple Representations
To succeed in STEM, students need two types of representational competencies; the ability to conceptually make sense of visual representations and perceptual fluency (similar to fluency in a language). We know little about how conceptual and perceptual competencies interact. Studies in chemistry courses at UW-Madison and Madison College address this question. Based on the results, we develop an educational technology that adapts to the individual's learning of these competencies: Chem Tutor. Funded by NSF DUE 1611782 (IUSE).

Modeling Perceptual Fluency with Visual Representations
STEM learning builds on perceptual fluency: the ability to “see” visual information implicitly and effortlessly. Common instruction focuses on explicit learning processes but cannot assess implicit processes that yield perceptual fluency and hence cannot adapt to students' perceptual fluency. To address this issue, studies test how students and experts perceive visual representations. Based on the results, we design perceptual fluency activities that use machine learning to adapt to students' needs. Funded by NSF IIS 1623605 (Cyberlearning).

Drawing Visual Representations
Educational technologies typically support conceptual learning processes by asking students to verbally explain concepts. When students work with visual representations, they may engage in visual rather than verbal processes. Drawing is an effective learning strategy, but we know little about how drawing helps students' learning with conventional representations. We investigate whether asking students to draw their own representations enhances the effectiveness of an educational technology that focuses on representational competencies.